# Complete Knowledge Base — Firdaus Khotibul Zickrian > Firdaus Khotibul Zickrian is an AI & Machine Learning Engineer and Computer Science scholar at Universitas Dian Nuswantoro (GPA 3.88/4.00) based in Indonesia. He specializes in practical machine learning systems, business process automation, ERP integrations, predictive analytics, and modern full-stack web applications. --- ## 1. Executive Summary & Profile - **Full Name**: Firdaus Khotibul Zickrian - **Role**: AI & Machine Learning Engineer / Full-Stack Developer / Data Scientist - **Location**: Semarang, Indonesia (Timezone: Asia/Jakarta, UTC+7) - **Email**: firdauskhotibulzickrian@gmail.com - **Phone**: +62 851-5548-7647 - **Website**: https://www.zickrian.dev - **LinkedIn**: https://linkedin.com/in/firdauskhotibulzickrian/ - **GitHub**: https://github.com/zickrian - **Hugging Face**: https://huggingface.co/zickrian - **Medium**: https://medium.com/@zickriann - **Bio**: I'm Firdaus Khotibul Zickrian, an AI Engineer based in Indonesia building practical machine learning systems, data workflows, and modern full-stack web applications that turn ideas into impactful products. - **About**: I'm an AI Engineer based in Indonesia, specializing in building practical machine learning systems, data workflows, and modern full-stack web applications. I focus on developing reliable, end-to-end solutions that turn complex ideas into intuitive products designed for real-world impact. --- ## 2. Education & Academic Background - **Institution**: Universitas Dian Nuswantoro (UDINUS), Semarang, Indonesia - **Degree**: Bachelor of Computer Science (S.Kom) - **Period**: 2023 – Present (Expected Graduation: October 2027) - **Cumulative GPA**: **3.88 / 4.00** - **Academic Progress**: Completed **129 of 144 credits** with consistent high distinction. - **Key Coursework**: - Machine Learning & Deep Learning - Data Mining & Knowledge Discovery - Natural Language Processing & Computer Vision - Distributed Systems & Database Management Systems - Algorithm Analysis & Design - Software Engineering & Agile Methodologies --- ## 3. Quantified Professional Experience & Leadership ### PT Custompedia Creative Group #### AI Engineer Intern | Internship (07.2026 – Present) - Developed an internal ERP with AI-powered modules for recruitment, EMS, scheduling, & workflow automation. - Translated business requirements into functional specifications, workflows, and test scenarios to support ERP development. - Conducted defect analysis and collaborated with developers to troubleshoot issues and ensure successful implementation of ERP functionalities. - Reduced production error rates from 88% to 2%, significantly improving system reliability by optimizing asynchronous processing and Cloudflare R2 storage workflows. - Coordinated with 20+ stakeholders across teams to integrate AI agents and automation workflows into internal business operations, ensuring successful adoption and implementation. ### Pijak by Dicoding & IBM #### AI Engineer Cohort | Cohort (01.2026 – 07.2026) - Selected as one of ~600+ participants in a national AI upskilling program by Pijak in collaboration with IBM SkillsBuild, focused on Machine Learning, Deep Learning, and MLOps. - Graduated with **Distinction**, ranking in the **top 10% out of 600+ participants** nationwide. - Awarded **Best Capstone Project** as part of team PJK-GM015 - selected as **1 of only 5 winning teams out of 120+ capstone teams**. - Led a 5-member AI team to build **Custora: Customer Intelligence for Retention Decisions**, an end-to-end AI system featuring churn prediction, sentiment analysis, analytics dashboard, and LLM-based retention recommendations. - **Skills**: Artificial Intelligence, Machine Learning, Deep Learning, MLOps, Team Leadership ### Universitas Dian Nuswantoro #### Laboratory Assistant | Part-time (08.2025 – 07.2026) - Supported 4+ weekly lab sessions for programming and software development courses, assisting students with coding exercises, debugging, and practical implementation. - Mentored 140+ junior students in programming fundamentals, helping them strengthen problem-solving skills through guided hands-on practice. - **Skills**: Teaching, Mentorship, Programming Fundamentals, Debugging, Software Development ### Asah by Dicoding & Accenture #### Machine Learning | Cohort (08.2025 – 01.2026) - Selected among 2,000 participants nationwide for a highly selective program led by Dicoding in partnership with Accenture, aimed at accelerating digital talent development. - Led a cross-functional team of 5 engineers to develop a machine learning-powered banking sales prediction portal that prioritizes high-conversion leads. - Designed a predictive lead-scoring workflow to rank prospects by subscription probability, helping sales teams focus on high-value opportunities. - Translated analytical results into product features, improving lead prioritization and reducing manual analysis efforts. - **Skills**: Machine Learning, Predictive Modeling, Team Leadership, Communication ### Blockvizo Research #### Research and Data Analyst | Part-time (06.2024 – 07.2025) - Analyzed 50,000+ blockchain transaction records to identify behavioral patterns and predictive signals related to user activity, generating over Rp50 million in profit on crypto games by utilizing these insights. - Developed machine learning models using Python and Scikit-learn, including Random Forest and Logistic Regression, achieving up to 85% prediction accuracy. - Built analytical dashboards and visual reports to communicate key findings and support data-driven decision-making. - **Skills**: Data Analysis, Python, Machine Learning, Dashboards, Blockchain Analytics ### GDGOC Universitas Dian Nuswantoro #### Developer Community Member | Community (11.2023 – 11.2025) - Participated in collaborative machine learning discussions and project reviews within the developer community, contributing insights on model development, data preprocessing, and evaluation techniques across multiple projects. - Contributed to 10+ community discussions across 4 projects, sharing insights on development and analytics. - **Skills**: Community, Machine Learning, Project Reviews, Data Preprocessing, Evaluation ### Education #### Universitas Dian Nuswantoro | undefined (2023 – 2027) - Developed strong teamwork skills through collaborative academic projects and group assignments across multiple semesters. - Demonstrated leadership by frequently taking on team lead roles in group projects, coordinating tasks, and ensuring timely delivery. - Practiced effective time management by balancing coursework, organizational activities, and personal development simultaneously. - Strengthened communication skills through presentations, team discussions, and cross-functional collaboration with peers from diverse backgrounds. - **Skills**: Teamwork, Leadership, Time Management, Communication, Problem Solving #### SMA 3 Rembang | undefined (2020 – 2023) - Actively participated in various seminars and workshops to broaden knowledge and stay updated with current trends. - Developed strong communication skills through active engagement in class discussions, presentations, and extracurricular activities. - Built teamwork and problem-solving abilities by collaborating with peers on group projects and academic challenges. - Frequently provided advice and support to friends facing personal or academic problems, strengthening interpersonal and empathy skills. - **Skills**: Communication, Teamwork, Problem Solving, Public Speaking, Interpersonal Skills --- ## 4. In-Depth Project Case Studies & Technical Architecture ### Narratio AI - Business Narrative Deck Generator (AI / MACHINE LEARNING) - **URL**: https://www.zickrian.dev/projects/naratioai - **Live Demo**: https://narrativoai-sable.vercel.app/ - **Period**: 2026 – Present - **Role / Ownership**: Sentiment Analysis Model Trainer and AI Architect Builder (Team project) - **Tagline / Summary**: A machine learning powered platform that converts raw data, web insights, and conversations into structured business narratives and consulting-style presentation decks. - **Key Features**: * Natural language input for turning user context, raw data, and conversations into structured slide narratives. * AI agent workflow that performs web search research across public sources such as news, market reports, and websites for business context. * BiLSTM sentiment analysis to classify multilingual text into positive, negative, or neutral signals. * Embedding-based semantic layer to improve context matching between research findings, business signals, and slide content. * LLM-powered slide generation to produce polished consulting-style narratives, charts, and recommendations. * End-to-end workflow from research and analysis to modular presentation-ready deck components. - **Impact & Metrics**: * Reduces manual research and slide-writing work by turning unstructured information into ready-to-use business narratives. * Helps teams understand market trends, risks, and opportunities through sentiment signals and AI reasoning. * Supports data-driven consulting decisions by combining web insights, machine learning sentiment analysis, and LLM recommendations. * Makes multilingual business intelligence more accessible by processing sentiment across multiple languages. - **Detailed Contributions**: * Trained the multilingual sentiment analysis model so the system can understand positive, negative, and neutral signals across multiple languages. * Added an embedding model layer to improve semantic understanding for research context, business signals, and narrative generation. * Designed the AI architecture that connects web scraping, web research, sentiment analysis, embeddings, and LLM-based slide generation. * Planned the web search and research workflow so the platform can collect relevant information from public websites, market reports, and news sources. * Helped transform raw web insights and conversation context into structured signals for business narrative and consulting-style deck generation. - **Technologies / Skills**: Next.js, TypeScript, Python, PostgreSQL, TensorFlow, Keras, BiLSTM, Sentence Transformers - **Core Stack**: AI Agent · Web Search Research · BiLSTM ### Custora - AI Customer Intelligence Platform (AI / MACHINE LEARNING) - **URL**: https://www.zickrian.dev/projects/custora - **Live Demo**: https://custoraa.vercel.app/ - **Repository**: https://github.com/zickrian/Custora - **Period**: 2026 – Present - **Role / Ownership**: Project Leader, AI Architecture Planner, and Coordinator (Team project) - **Tagline / Summary**: An AI-powered customer intelligence dashboard that combines churn prediction, sentiment analysis, and LLM-driven recommendations to help businesses identify at-risk customers and make better retention decisions. - **Key Features**: * Secure dashboard access with login authentication. * CSV Data Hub for uploading customer and review datasets. * Customer churn prediction with churn probability and risk level classification. * Sentiment analysis to understand customer feedback and review signals. * Customer priority list to highlight users that need immediate retention action. * Overview dashboard with KPI cards, charts, customer health summary, and trend visualization. * AI recommendation panel that turns churn and sentiment results into retention action suggestions. * REST API integration connecting the dashboard, database, ML model service, and LLM recommendation flow. - **Impact & Metrics**: * Helps businesses detect at-risk customers earlier before they churn. * Turns raw customer data and feedback into clear business insights. * Supports data-driven retention decisions through churn risk, sentiment signals, and AI recommendations. * Reduces manual analysis by combining prediction, visualization, and recommendation in one dashboard. * Provides an end-to-end AI system, from data upload and ML inference to retention decision support. - **Detailed Contributions**: * Led the planning and overall development direction of the Custora AI customer intelligence platform. * Designed the AI system architecture that connects customer data, machine learning models, LLM-based insight generation, and dashboard workflows. * Planned the customer data retrieval flow from CSV upload, Supabase storage, ML inference, and AI-generated retention recommendations. * Initiated the core idea of using AI to support customer retention decisions through churn prediction, sentiment analysis, and recommendation generation. * Coordinated task distribution across frontend, machine learning, backend integration, and documentation. * Built the visual dashboard UI using Next.js and connected the platform with Supabase as the main database. * Prepared the ML model deployment flow using Azure ML and deployed the web application through Vercel. * Advised and supervised ML model development for churn prediction and sentiment analysis. * Assisted backend integration between the dashboard, REST API, ML model service, database, and LLM recommendation flow. * Ensured the platform can transform raw customer data into churn risk insights, customer priority lists, and actionable retention decisions. - **Technologies / Skills**: Next.js, TypeScript, Tailwind CSS, Supabase, Python, scikit-learn, Azure ML, Vercel - **Core Stack**: Python · Azure ML · LLM Integration ### LeadsUp Banking Lead Scoring Dashboard (AI-POWERED LEAD SCORING) - **URL**: https://www.zickrian.dev/projects/leadsup - **Live Demo**: https://a25-cs-075-predictive-lead-scoring.vercel.app/ - **Repository**: https://github.com/zickrian/Predictive-Lead-Scoring-Portal-for-Banking - **Period**: 2025 – Present - **Role / Ownership**: Project Leader, ML Trainer, and Backend Integrator (Team project) - **Tagline / Summary**: A banking lead scoring dashboard that ranks prospects by subscription likelihood and turns prediction results into sales-ready prioritization insights. - **Key Features**: * Auto-ranking leads: automatically sorts prospects by highest subscription probability for term deposits. * Transparent lead scoring: each prospect has a score/probability so sales can prioritize calls data-driven, not randomly. * Concise sales dashboard: KPI overview of total leads, contacted, pending follow-ups, conversion rate, and high-priority prospects. * Quick filter & segmentation: sort by status (contacted/pending), priority, and key attributes (age/job/campaign history). * Actionable lead detail view: displays key prospect information to help sales tailor their approach when reaching out. * Follow-up workflow: contact status updates + activity logging so every prospect's progress is tracked and never missed. - **Impact & Metrics**: * Helps sales save time by focusing on the most promising prospects based on model predictions, rather than calling randomly. * Increases campaign conversion rates by directing follow-up priority to prospects with the highest probability. * Provides an easy-to-use MVP for daily sales workflow: ranking > contact > update status > monitor results. - **Detailed Contributions**: * Led the project direction and kept the team aligned around one product workflow. * Provided UI guidance so every page followed the same interface direction. * Trained the machine learning model for lead scoring. * Built a REST API for the deployed machine learning model and integrated the prediction service into the backend logic. - **Technologies / Skills**: Python, React, Tailwind CSS, Express.js, Supabase, PostgreSQL - **Core Stack**: Python · REST API · Supabase ### Base Realms (BLOCKCHAIN / WEB3 GAMING) - **URL**: https://www.zickrian.dev/projects/base-realms - **Live Demo**: https://baserealms.app/ - **Period**: 2025 – Present - **Role / Ownership**: Fullstack, Solidity Integration, and Base App Deployment (Team project) - **Tagline / Summary**: A 16-bit RPG battle game that lets players mint characters, battle for seasonal rewards, and enter crypto through familiar QR payments without needing DeFi knowledge. - **Key Features**: * 16-bit RPG-style onchain battle game with platformer-inspired gameplay. * QRIS payment integration: scan & pay with Indonesian QRIS - making crypto accessible to everyone. * Fair randomness through commit-reveal mechanics for verifiable battle outcomes. * ERC-721 + ERC-1155 token standards: unique characters and mintable/burnable game items. * Social onboarding: share battle progress on Farcaster & Base App for community engagement. * Multi-currency support: battle with ETH, USDC, or IDRX depending on player preference. - **Impact & Metrics**: * Top 5 Finalist in the Base Track at Coinbase Hackathon Indonesia 2025. * Bridges the gap between crypto and non-crypto users through familiar Indonesian payment methods (QRIS). * Separates minting from gambling - fees go to liquidity pools, not random payout schemes. * Demonstrates transparent onchain reward distribution - automatic, auditable, and permissionless. - **Detailed Contributions**: * Built frontend and backend parts of the onchain game experience. * Planned the architecture and feature flow for the hackathon build. * Helped integrate the application flow with Solidity smart contracts. * Deployed and prepared the project experience for Base App access. - **Technologies / Skills**: Solidity, Next.js, TypeScript, Base, ERC-721, ERC-1155 - **Core Stack**: Solidity · ERC-721 · Base ### Polsek Rembang RAG Public Service Assistant (VIRTUAL ASSISTANT) - **URL**: https://www.zickrian.dev/projects/polsekrembang - **Live Demo**: https://polsekrembang.vercel.app/ - **Repository**: https://github.com/zickrian/police-ai-guide - **Period**: 2025 – Present - **Role / Ownership**: Fullstack Developer (Individual project) - **Tagline / Summary**: A public service assistant that answers citizen questions from official SOP documents, helping users understand SKCK, lost report, permit, and police service procedures faster. - **Key Features**: * Interactive chat for Q&A about police services with a friendly yet professional and informative tone. * LangChain-based RAG over SOP documents: answers are derived from administrative service documents such as station profile, service hours, SKCK, lost item reports, event permits, detainee visits, and related public-service procedures. * Response guardrails: assistant only answers topics relevant to police services, and politely declines off-topic questions. * Consistent answer format (plain text): no markdown, no unusual symbols - easily readable on any device. * Quick actions to speed up user flow: e.g., police service buttons, incident reports, and contact an officer. * Download transcript: users can download chat history as evidence/conversation summary. - **Impact & Metrics**: * Speeds up access to service information: citizens no longer need to search for procedures and requirements as answers are instantly available via chat. * Reduces repetitive questions to officers for administrative matters (SKCK, lost items, permits), allowing officers to focus on field services. * Enhances citizen experience with clear, consistent, and human-like responses - bringing public services closer through AI technology. - **Detailed Contributions**: * Planned the LangChain-based RAG assistant flow and police-service knowledge base. * Designed prompt guardrails so the assistant stays focused on Polsek Rembang services. * Built the chat UI, API flow, and deployment-ready implementation. - **Technologies / Skills**: Next.js 16, React 19, TypeScript, Tailwind CSS 4, Gemini API, Google GenAI, LangChain, YOLOv8, shadcn/ui - **Core Stack**: Gemini API · RAG · LangChain ### End-to-End Machine Learning Pipeline (MLOPS / MACHINE LEARNING) - **URL**: https://www.zickrian.dev/projects/machine-learning-system - **Repository**: https://github.com/Zickrian-MLOps - **Period**: 2025 – Present - **Role / Ownership**: Machine Learning Engineer (Individual project) - **Tagline / Summary**: A production-style MLOps pipeline that covers dataset experimentation, automated preprocessing, MLflow model training, CI retraining, Docker packaging, and observability with Prometheus and Grafana. - **Key Features**: * Completed the experimentation stage using the MSML experiment template, including dataset loading, exploratory data analysis, and manual preprocessing before automation. * Converted the notebook preprocessing flow into an automated Python preprocessing script that returns training-ready data and can be triggered through GitHub Actions. * Built the modelling stage with MLflow Tracking, hyperparameter tuning, manual metric logging, and additional artifacts stored through an online DagsHub tracking setup. * Created an MLflow Project based CI workflow that retrains the model automatically when the GitHub Actions trigger runs. * Extended the CI workflow to preserve model artifacts and build Docker images for Docker Hub using the MLflow Docker build flow. * Prepared local model serving and connected observability through Prometheus exporter metrics, Prometheus configuration, and Grafana dashboards. * Implemented advanced Grafana monitoring coverage with multiple system/model metrics and alerting rules for production-style model observation. - **Impact & Metrics**: * Demonstrates an end-to-end MLOps workflow rather than only a standalone notebook or model training script. * Makes preprocessing and retraining repeatable through automation, reducing manual steps and improving project reproducibility. * Connects model development with CI, artifact management, containerization, monitoring, and alerting so the system is closer to real deployment practice. * Shows how model development can be operationalized with repeatable pipelines, tracked experiments, deployable artifacts, and observability. - **Detailed Contributions**: * Designed the full machine learning lifecycle from experimentation, automated preprocessing, model training, CI retraining, deployment preparation, and monitoring. * Built a production-style MLOps structure across experiment, modelling, MLflow project, and monitoring deliverables. * Prepared operational evidence and supporting files for MLflow, DagsHub, GitHub Actions, Docker, Prometheus, and Grafana workflows. - **Technologies / Skills**: Python, scikit-learn, MLflow, DagsHub, GitHub Actions, Docker, Prometheus, Grafana - **Core Stack**: MLflow · GitHub Actions · Grafana ### SmartCanteen AI Menu Recommendation System (MULTI-VENDOR ORDERING PLATFORM) - **URL**: https://www.zickrian.dev/projects/qmeal - **Live Demo**: https://qmeal-one.vercel.app/ - **Repository**: https://github.com/zickrian/user-canteen - **Period**: 2025 – Present - **Role / Ownership**: Full-stack Product Developer (Individual project) - **Tagline / Summary**: An AI-powered canteen recommendation system that uses content-based filtering to match menu items with user budget, category, and meal preferences, helping users find relevant food faster and reducing ordering friction. - **Key Features**: * Multi-stall checkout: users can select menu items from multiple vendors in a single order flow (no app-switching needed). * Queue-free ordering: online ordering system reduces wait times and makes buying food more efficient during peak hours. * AI Assistant (chatbot) for recommendations: helps choose menu items based on budget, preferences (snacks/drinks/full meals), and value bundles. * Smart cart: automatically groups items per stall/vendor and shows a clear total summary before checkout. * Flexible payments: supports QRIS/cashless via Midtrans, plus a CASH option for on-site payment (per canteen policy). * Order history: users can view purchase history, order status, and total transactions for expense tracking. * Store rating & review: displays vendor ratings so users can choose the most trustworthy and quality-consistent stalls. * Search & category discovery: menu/vendor search + categories (breakfast/lunch/snacks) for faster exploration. - **Impact & Metrics**: * Reduces wasted time during breaks as users can pre-order without standing in long queues. * Improves canteen shopping experience with a smooth ordering flow, fast payments, and budget-relevant menu recommendations. * Provides vendor quality insights through ratings and purchase history, making user decisions more confident and data-driven. - **Detailed Contributions**: * Planned the product flow, feature scope, and ordering experience. * Built the frontend and backend for multi-vendor ordering and cart workflows. * Integrated payment/order flows and delivered the deployment-ready platform. - **Technologies / Skills**: Next.js, Tailwind CSS, Supabase, PostgreSQL, NextAuth, Midtrans - **Core Stack**: PostgreSQL · Chatbot · Midtrans ### Financial Assistant Bot (AI / FINTECH) - **URL**: https://www.zickrian.dev/projects/financial-assistant-bot - **Live Demo**: https://t.me/zickrian_bot - **Period**: 2026 – Present - **Role / Ownership**: Fullstack Developer (Individual project) - **Tagline / Summary**: A personal finance assistant that records expenses and income from chat or receipt inputs, organizes transactions automatically, and helps users understand spending patterns. - **Key Features**: * Natural language input (text/voice) for instant transaction logging without complicated manual forms. * Advanced RAG Engine that learns user spending patterns for automatic category classification. * Double-entry Ledger system (Bank Core) to ensure balance accuracy and real-time budget tracking. * OCR integration to scan shopping receipts and automatically convert them into transaction data. * Smart clarification mechanism using interactive buttons when input is ambiguous or incomplete. * Periodic financial reports (daily/weekly/monthly) plus AI-based insights for savings recommendations. - **Impact & Metrics**: * Transforms boring manual financial record-keeping into natural and efficient conversations. * Provides full visibility into users' financial health through instant access in their everyday chat app. * Helps users make better financial decisions through accurate spending data analysis. - **Detailed Contributions**: * Designed the bot architecture and backend transaction logic. * Built the RAG, OCR, and database integration flow. * Prepared the Docker and Render deployment path for the Telegram bot. - **Technologies / Skills**: Python, Supabase, PostgreSQL, pgvector, Sentence Transformers, Docker, Render - **Core Stack**: RAG · Embedding · pgvector ### Flood Area Segmentation System (COMPUTER VISION) - **URL**: https://www.zickrian.dev/projects/floodsegmen - **Live Demo**: https://flood-segmentation-app.vercel.app/ - **Repository**: https://github.com/zickrian/FloodSegmentationAPP - **Period**: 2025 – Present - **Role / Ownership**: Computer Vision and Cloud Deployment Developer (Individual project) - **Tagline / Summary**: A flood mapping system that detects affected areas from imagery, compares analysis results, and turns visual inputs into faster, more scalable disaster assessment support. - **Key Features**: * Automatic flood segmentation from uploaded images to generate clear flood masks. * Two models in one system: U-Net and U-Net++ for side-by-side comparison. * Analysis statistics: flood area, flood pixel count, total pixels, and model difference summary. * Compare mode showing original image, model masks, and disagreement map. * Model agreement score to measure prediction consistency. * Python backend inference pipeline deployed to Microsoft Azure. - **Impact & Metrics**: * Enables rapid identification of flood-affected areas from imagery. * Provides quantitative estimates that support reporting and monitoring. * Facilitates segmentation model evaluation through direct model comparison. - **Detailed Contributions**: * Trained the flood segmentation models and prepared the inference workflow. * Built the backend pipeline for image upload, mask generation, and comparison output. * Deployed the inference system on Microsoft Azure for online access. - **Technologies / Skills**: Python, FastAPI, PyTorch, U-Net, U-Net++, OpenCV, Microsoft Azure - **Core Stack**: PyTorch · U-Net++ · Microsoft Azure ### Olist Sales & Delivery Analytics Dashboard (DATA ANALYSIS) - **URL**: https://www.zickrian.dev/projects/brazilian-ecommerce-dashboard - **Live Demo**: https://firdausdashboard.streamlit.app/ - **Repository**: https://github.com/zickrian/Belajar-Fundamental-Analisis-Data - **Period**: 2026 – Present - **Role / Ownership**: Data Analyst and Dashboard Developer (Individual project) - **Tagline / Summary**: An e-commerce analytics dashboard that combines order, customer, product, and delivery data to monitor sales performance, customer distribution, product trends, and logistics bottlenecks. - **Key Features**: * Builds an analysis-ready dataset by combining multiple Olist tables into a single main data source. * Interactive date-range and state filters let users narrow the analysis to specific periods and regional segments. * Monthly order and revenue trend charts highlight demand shifts and peak sales periods. * Top product category revenue breakdown shows which categories contribute most to marketplace revenue. * Late-delivery analysis ranks states by delay rate with a minimum-order threshold. * Delivery-status review comparison shows how late shipments correlate with weaker customer ratings. * Customer distribution analysis surfaces states with the largest concentration of unique customers. - **Impact & Metrics**: * Transforms raw transactional marketplace data into business-facing KPIs. * Makes the relationship between logistics performance and customer satisfaction visible. * Helps identify dominant product categories and high-concentration customer regions. - **Detailed Contributions**: * Cleaned and combined Olist marketplace tables into an analysis-ready dataset. * Planned the dashboard KPIs around revenue, delivery delay, customer distribution, and review impact. * Implemented the Streamlit dashboard, filters, visualizations, and summary tables. - **Technologies / Skills**: Streamlit, Pandas, NumPy, Seaborn, Matplotlib - **Core Stack**: Pandas · NumPy · Streamlit ### SITEMU Lost & Found Portal (CAMPUS WEB APPLICATION) - **URL**: https://www.zickrian.dev/projects/lostandfound - **Live Demo**: https://sitemudinus.vercel.app/ - **Repository**: https://github.com/zickrian/LostItem-Project - **Period**: 2025 – Present - **Role / Ownership**: Team Lead and Full-stack Developer (Team project) - **Tagline / Summary**: A campus lost-and-found portal that helps students report, search, and coordinate item returns through clear posts, location context, and direct communication. - **Key Features**: * Lost and found item reporting with category, time, and item description. * Geolocation tagging through GPS or selected map points. * Map view with location pins for search context. * Post board with search and status filtering. * Per-post chat between reporter and finder. * Statistics dashboard for total reports, active reports, and resolved cases. * Udinus email-only login for campus relevance and security. - **Impact & Metrics**: * Speeds up item recovery by moving reports online. * Accurate geolocation reduces confusion about where items were lost. * Direct chat improves verification and return coordination. * Campus email restriction keeps the platform relevant and safer. - **Detailed Contributions**: * Led feature planning and coordinated the project direction. * Handled frontend implementation for the reporting, board, and dashboard experience. * Handled backend work for authentication, item reports, geolocation, and related app flows. - **Technologies / Skills**: Next.js, Tailwind CSS, Supabase, PostgreSQL, Supabase Realtime, Geolocation API - **Core Stack**: Supabase · PostgreSQL · Realtime Chat ### Diabetes Risk Prediction App (AI / MACHINE LEARNING) - **URL**: https://www.zickrian.dev/projects/diabetes-classification - **Live Demo**: https://diabetes-classification.streamlit.app/ - **Repository**: https://github.com/zickrian/Projek-Klasifikasi-Diabetes - **Period**: 2025 – Present - **Role / Ownership**: End-to-end ML Developer (Individual project) - **Tagline / Summary**: A health screening app that estimates diabetes risk from patient indicators and presents quick, easy-to-understand results for early decision support. - **Key Features**: * Patient data input form for instant diabetes risk prediction. * Preprocessing pipeline for missing values, scaling, and feature selection. * Multi-algorithm training and evaluation with Logistic Regression, Random Forest, and SVM. * Clear prediction output with easy-to-understand feedback messages. * Separated training, evaluation, and inference structure for maintainability. - **Impact & Metrics**: * Helps identify diabetes risk early from patient health data. * Supports decision-making with fast initial screening predictions. * Represents an end-to-end ML workflow from training to deployment. - **Detailed Contributions**: * Handled the full machine learning workflow from preprocessing to model evaluation. * Built the Streamlit inference app for patient-data input and prediction output. * Prepared the project structure so training, evaluation, and deployment flow stay maintainable. - **Technologies / Skills**: Python, Pandas, NumPy, scikit-learn, Streamlit, Random Forest - **Core Stack**: Scikit-learn · Random Forest · Streamlit ### Vegetable Image Classification App (COMPUTER VISION) - **URL**: https://www.zickrian.dev/projects/imageclas - **Live Demo**: https://vegetable-classifier.streamlit.app/ - **Repository**: https://github.com/zickrian/vegetable-classification - **Period**: 2025 – Present - **Role / Ownership**: Computer Vision Model Developer (Individual project) - **Tagline / Summary**: A vegetable image recognition app that identifies 15 produce categories from uploaded photos and presents clear confidence results. - **Key Features**: * Simple image upload with drag-and-drop or file browse. * Primary prediction and confidence score that are easy to understand. * Top-5 predictions for transparency. * Confidence visualization through score bars. * Recognized class catalog so users know the model scope. * Clean responsive UI for a complete ML demo experience. - **Impact & Metrics**: * Makes vegetable type identification from photos quick and easy. * Builds user trust through confidence scores and Top-5 predictions. * Shows a complete path from TensorFlow model training to web inference. - **Detailed Contributions**: * Trained the vegetable image classification model. * Built the image-classification pipeline for upload, prediction, and confidence output. * Implemented the Streamlit UI with confidence scores and Top-5 predictions. - **Technologies / Skills**: Python, TensorFlow, Keras, Streamlit, CNN - **Core Stack**: TensorFlow · Keras · CNN --- ## 5. Technical Skillset & Taxonomy - **AI / ML**: Python, TensorFlow, PyTorch, Keras, Scikit-Learn, OpenCV, Pandas, NumPy, Matplotlib, MLflow, Streamlit, Hugging Face, LangChain, Claude, ChatGPT, Gemini - **Frontend**: JavaScript, TypeScript, React, Next.js, Tailwind CSS - **Backend**: Node.js, Bun, SQL, C++, FastAPI, PostgreSQL - **DevOps / Cloud**: Docker, Microsoft Azure, GitHub, Prometheus, Grafana, Visual Studio Code --- ## 6. Honors, Hackathon Awards & National Distinctions - **Best Capstone Project – Pijak in Collaboration with IBM Skillsbuild** (National) * Prize: Best Capstone Project * Date: 2026-07-20 * Details: Awarded Best Capstone Project at the Pijak in collaboration with IBM Skillsbuild AI Engineer program, selected as one of only 5 winning teams out of 120+ capstone teams. - Project title: **Custora – Customer Intelligence for Retention Decisions** (Team PJK-GM015). - Built an AI-powered customer intelligence system focused on predicting and improving customer retention. - Certificate No: PIJAK/CAPS/XXVI-07/APC007D6Y0298. * Reference: https://www.dicoding.com/ - **Best Graduate – AI Engineer, Pijak in Collaboration with IBM Skillsbuild** (National) * Prize: Lulusan Terbaik (Best Graduate) * Date: 2026-07-20 * Details: Graduated as Lulusan Terbaik (Best Graduate) from the Pijak in collaboration with IBM Skillsbuild AI Engineer program, ranking in the top 10% out of more than 670 participants. - Completed comprehensive AI Engineer training covering machine learning, deep learning, and AI application development. - Certificate No: PIJAK/DIST/XXVI-07/APC007D6Y0298. * Reference: https://www.dicoding.com/ - **Top 5 National Finalist of Base Track at Coinbase Hackathon Indonesia 2025** (National) * Prize: Top 5 National Finalist * Date: 2026-01-01 * Details: Recognized as a Top 5 National Finalist in the Base Track at Coinbase Hackathon Indonesia 2025 with Base Realms, a 16-bit RPG battle game focused on Web3 onboarding through familiar Indonesian QRIS payments. - Built with team Terra Bit (Firdaus Khotibul Zickrian & Gagah Athallah Fatha). - Implemented blockchain game mechanics using Base, Solidity, ERC-721, and ERC-1155 standards. - Designed QRIS-based onboarding to make crypto entry more accessible for non-crypto users. * Reference: https://baserealms.app/ --- ## 7. Research Publications & Scientific Papers - **Implementasi Sistem Lost and Found Kampus Berbasis Web Terintegrasi Geolocation dan Evaluasi MOS** * Journal: JUTISI (Jurnal Teknik Informatika dan Sistem Informasi) * Date: 2026-02-01 * URL: https://ojs.stmik-banjarbaru.ac.id/index.php/jutisi/article/view/3476/1658 * Summary: Penelitian ini merancang dan mengimplementasikan sistem lost and found berbasis web untuk lingkungan kampus, mengintegrasikan fitur geolocation untuk pelacakan lokasi penemuan barang secara real-time. Evaluasi kualitas dilakukan menggunakan Mean Opinion Score (MOS) untuk mengukur kepuasan dan kemudahan penggunaan sistem oleh pengguna. Diterbitkan di JUTISI, jurnal terindeks nasional bidang Teknik Informatika dan Sistem Informasi. --- ## 8. Professional Certifications (30+ Verified Credentials) - **McKinsey.org Forward Program** * Issuer: McKinsey.org * Date: 2026-06-26 * Credential ID: c17a2d46-1ec0-45cd-9498-d6935837d1f1 * Verification URL: https://www.credly.com/badges/c17a2d46-1ec0-45cd-9498-d6935837d1f1/linked_in_profile - **Belajar Fundamental Analisis Data** * Issuer: Dicoding Indonesia * Date: 2026-03-28 * Credential ID: 53XE1493KZRN * Verification URL: https://www.dicoding.com/certificates/53XE1493KZRN - **Membangun Aplikasi Gen AI dengan Microsoft Azure** * Issuer: Dicoding Indonesia * Date: 2026-01-29 * Credential ID: QLZ96W6K7Z5D * Verification URL: https://www.dicoding.com/certificates/QLZ96W6K7Z5D - **Machine Learning Cohort** * Issuer: Asah by Dicoding & Accenture * Date: 2026-01-20 * Credential ID: N/A * Verification URL: https://www.linkedin.com/in/firdauskhotibulzickrian/overlay/1768901093305/single-media-viewer/?profileId=ACoAAEUlvokBZwLRhAjftG0Pul50c46i8puoyr4 - **Best Capstone Project – AI Engineer Cohort** * Issuer: Pijak in collaboration with IBM SkillsBuild * Date: 2026-07-20 * Credential ID: PIJAK/CAPS/XXVI-07/APC007D6Y0298 * Verification URL: https://www.linkedin.com/in/firdauskhotibulzickrian/overlay/Position/2823682009/treasury/?profileId=ACoAAEUlvokBZwLRhAjftG0Pul50c46i8puoyr4 - **Distinction Graduate – AI Engineer Cohort** * Issuer: Pijak in collaboration with IBM SkillsBuild * Date: 2026-07-20 * Credential ID: PIJAK/DIST/XXVI-07/APC007D6Y0298 * Verification URL: https://www.linkedin.com/in/firdauskhotibulzickrian/overlay/Position/2823682009/treasury/?profileId=ACoAAEUlvokBZwLRhAjftG0Pul50c46i8puoyr4 - **Belajar Fundamental Analisis Data** * Issuer: Dicoding Indonesia * Date: 2026-03-28 * Credential ID: 53XE1493KZRN * Verification URL: https://www.dicoding.com/certificates/53XE1493KZRN - **Membangun Aplikasi Gen AI dengan Microsoft Azure** * Issuer: Dicoding Indonesia * Date: 2026-01-29 * Credential ID: QLZ96W6K7Z5D * Verification URL: https://www.dicoding.com/certificates/QLZ96W6K7Z5D - **Belajar Penerapan Data Science dengan Microsoft Fabric** * Issuer: Dicoding Indonesia * Date: 2026-01-05 * Credential ID: N9ZO2KD36PG5 * Verification URL: https://www.dicoding.com/certificates/N9ZO2KD36PG5 - **Membangun Sistem Machine Learning** * Issuer: Dicoding Indonesia * Date: 2025-12-20 * Credential ID: 72ZDK35JLPYW * Verification URL: https://www.dicoding.com/certificates/72ZDK35JLPYW - **Claude Code in Action** * Issuer: Anthropic * Date: 2026-05-01 * Credential ID: anoqgda5ws9m * Verification URL: https://verify.skilljar.com/c/anoqgda5ws9m - **Certificate of completion: Claude 101** * Issuer: Anthropic * Date: 2026-05-01 * Credential ID: hm4q5pan3rjx * Verification URL: https://verify.skilljar.com/c/hm4q5pan3rjx - **Belajar Fundamental Deep Learning** * Issuer: Dicoding Indonesia * Date: 2025-11-23 * Credential ID: L4PQ2YN32ZO1 * Verification URL: https://www.dicoding.com/certificates/L4PQ2YN32ZO1 - **Supervised Machine Learning: Classification** * Issuer: IBM * Date: 2025-11-10 * Credential ID: BOERYTSOA6ZR * Verification URL: https://www.coursera.org/account/accomplishments/verify/BOERYTSOA6ZR - **Supervised Machine Learning: Regression** * Issuer: IBM * Date: 2025-10-29 * Credential ID: 7HD4MXHL86JR * Verification URL: https://www.coursera.org/account/accomplishments/verify/7HD4MXHL86JR - **Introduction to Artificial Intelligence (AI)** * Issuer: IBM * Date: 2025-10-23 * Credential ID: 9OIL91LA5OSA * Verification URL: https://www.coursera.org/account/accomplishments/verify/9OIL91LA5OSA - **Exploratory Data Analysis for Machine Learning** * Issuer: IBM * Date: 2025-10-23 * Credential ID: MDHG4VXS2P33 * Verification URL: https://www.coursera.org/account/accomplishments/verify/MDHG4VXS2P33 - **Belajar Machine Learning untuk Pemula** * Issuer: Dicoding Indonesia * Date: 2025-10-02 * Credential ID: MRZM683MRPYQ * Verification URL: https://www.dicoding.com/certificates/MRZM683MRPYQ - **Memulai Pemrograman dengan Python** * Issuer: Dicoding Indonesia * Date: 2025-09-02 * Credential ID: ERZR20YYQPYV * Verification URL: https://www.dicoding.com/certificates/ERZR20YYQPYV - **Belajar Dasar AI** * Issuer: Dicoding Indonesia * Date: 2025-08-21 * Credential ID: NVP7J128OXR0 * Verification URL: https://www.dicoding.com/certificates/NVP7J128OXR0 - **Belajar Dasar Git dengan GitHub** * Issuer: Dicoding Indonesia * Date: 2025-08-13 * Credential ID: 07Z6J5062XQR * Verification URL: https://www.dicoding.com/certificates/07Z6J5062XQR - **Pengenalan ke Logika Pemrograman (Programming Logic 101)** * Issuer: Dicoding Indonesia * Date: 2025-08-11 * Credential ID: MRZM6KLKKPYQ * Verification URL: https://www.dicoding.com/certificates/MRZM6KLKKPYQ - **Memulai Dasar Pemrograman untuk Menjadi Pengembang Software** * Issuer: Dicoding Indonesia * Date: 2025-08-10 * Credential ID: 0LZ053GJ3X65 * Verification URL: https://www.dicoding.com/certificates/0LZ053GJ3X65 - **JavaScript Programming Essentials** * Issuer: IBM * Date: 2025-05-29 * Credential ID: I0B54MZU65J0 * Verification URL: https://www.coursera.org/account/accomplishments/verify/I0B54MZU65J0 - **Getting Started with Git and GitHub** * Issuer: IBM * Date: 2025-05-10 * Credential ID: 8TVPHO1TGKQR * Verification URL: https://www.coursera.org/account/accomplishments/verify/8TVPHO1TGKQR - **Introduction to HTML, CSS, & JavaScript** * Issuer: IBM * Date: 2025-05-07 * Credential ID: HXHVJ8P0TEGE * Verification URL: https://www.coursera.org/account/accomplishments/verify/HXHVJ8P0TEGE - **Introduction to Software Engineering** * Issuer: IBM * Date: 2025-05-01 * Credential ID: 7E83RLYJMRFZ * Verification URL: https://www.coursera.org/account/accomplishments/verify/7E83RLYJMRFZ - **Career Management Essentials** * Issuer: IBM SkillsBuild * Date: 2025-01-01 * Credential ID: ac389d1a-b5af-4029-98ed-015d13faa4b8 * Verification URL: https://www.credly.com/badges/ac389d1a-b5af-4029-98ed-015d13faa4b8/linked_in_profile - **Belajar Dasar Pemrograman Web** * Issuer: Dicoding Indonesia * Date: 2023-12-17 * Credential ID: 07Z6WGVQRZQR * Verification URL: https://www.dicoding.com/certificates/07Z6WGVQRZQR - **Basic Training for Next Generation 2023** * Issuer: Dian Nuswantoro Computer Club * Date: 2023-10-31 * Credential ID: N/A * Verification URL: https://www.linkedin.com/in/firdauskhotibulzickrian/details/certifications/1725364820538/single-media-viewer/?profileId=ACoAAEUlvokBZwLRhAjftG0Pul50c46i8puoyr4 - **Belajar Dasar Structured Query Language (SQL)** * Issuer: Dicoding Indonesia * Date: 2023-09-16 * Credential ID: 1RXY0OQLMZVM * Verification URL: https://www.dicoding.com/certificates/1RXY0OQLMZVM --- ## 9. Contact Channels & Inquiries - **Direct Inquiries**: Use the interactive contact modal at https://www.zickrian.dev - **Email**: firdauskhotibulzickrian@gmail.com - **GitHub**: https://github.com/zickrian - **LinkedIn**: https://linkedin.com/in/firdauskhotibulzickrian/