Fiqih Fathor Rachim
AI Engineer | Production AI/ML | Document Intelligence • NLP • Computer Vision | PyTorch
Specializing in Document Intelligence & NLP for Financial Services AI Engineer
LinkedIn Email
About Me
AI Engineer with 2+ years building production machine learning systems
for financial services at Bank Mega. Specializing in Document Intelligence,
NLP, and Computer Vision solutions that process 10,000+ documents daily
across 6+ Indonesian banks.
Currently expanding cloud-native ML expertise through AWS Machine Learning
Specialty certification to architect scalable, production-ready AI systems.
Key Expertise:
- Production ML deployment (LLM, Computer Vision, NLP)
- Document Intelligence & OCR systems
- Cloud-native ML architecture (AWS - certifying)
- MLOps and API development (FastAPI, Docker)
- Financial services domain knowledge
Skills
AI/ML Frameworks
- PyTorch, TensorFlow, Scikit-Learn, Hugging Face Transformers
- Large Language Models (LangChain, LangGraph, n8n workflow automation)
- Computer Vision (EfficientNet, ResNet, Vision Transformers)
- NLP (BERT, GPT, Indonesian language models)
Cloud & MLOps
- Docker
- MLflow, CI/CD pipelines
- Model monitoring and optimization
Backend Development
- Python (FastAPI, Flask)
- PostgreSQL, Redis
- API design and microservices
- Real-time data processing (Kafka)
Programming Languages
- Python (Expert), SQL, Rust
Education
Bachelor of Science, Mathematics
Universitas Airlangga | 2017 - 2022
Focus Areas: Deep Learning, Natural Language Processing, Statistical Methods
Thesis: Text Classification using Hybrid Convolutional Neural Network
and Gated Recurrent Unit methods
Work Experience
AI Engineer @ Bank Mega
February 2023 - Present | Jakarta, Indonesia
Building production AI/ML systems for financial document intelligence and
automation, serving 6+ major Indonesian banks with 10,000+ daily document
processing.
Key Projects & Impact:
🏦 AI-Powered Credit Control Monitoring
- Developed LLM-based monitoring system with n8n workflow automation
- Processes multiple document categories with significant efficiency improvement
- Tech: n8n, FastAPI, PostgreSQL, Redis, GPT-4, Mistral OCR
- Impact: 40% reduction in assessment time, improved fraud detection accuracy
🏢 Multi-Bank Statement Processing System
- Built intelligent extraction system supporting 6 major Indonesian banks
- Advanced table detection and regex-based parsing
- Tech: Table Transformer, Regex, Pandas, PostgreSQL
- Impact: 60% reduction in manual processing time
📄 Enterprise OCR System for KTP Processing
- ID card processing pipeline using TrOCR and CRAFT frameworks
- High-volume document processing with significant reduction in manual verification
- Tech: PyTorch, TrOCR, CRAFT, OpenCV, FastAPI, Docker
- Impact: Processing 1000+ documents/day with 95%+ accuracy
💰 Financial Document Classification
- Image-based classification using EfficientNet, ResNet, Vision Transformers
- Integrated with FastAPI for API interactions, tracked with MLflow
- Tech: PyTorch, ViT, FastAPI, MLflow
- Impact: Automated categorization of financial documents
Featured Projects
💼 Financial Document Classification
GitHub Repository
Documentation
Deep learning system for automated financial document categorization using
Vision Transformers, EfficientNet, and ResNet architectures.
Key Features:
- Multi-model ensemble for 95%+ accuracy
- FastAPI REST API with <100ms inference time
- MLflow integration for experiment tracking
- Production-ready deployment with Docker
Tech Stack: PyTorch, Vision Transformers, FastAPI, MLflow, Docker- OCR System for Document Recognition
Designed and deployed an OCR system to recognize and extract text from identity cards and financial documents.
📝 OCR System for Document Recognition
Deployed OCR system for identity card and financial document text extraction.
Key Features:
- TrOCR and CRAFT frameworks for accurate text detection
- Real-time processing pipeline
- Support for multiple document types
- Batch processing capabilities
Tech Stack: PyTorch, TrOCR, CRAFT, OpenCV, FastAPI, Docker
Impact: Processes 1000+ documents daily with 95%+ accuracy
📊 Financial Dashboard
Interactive dashboards providing real-time insights into financial data for
stakeholder decision-making.
Key Features:
- Real-time data visualization
- Custom analytics and reporting
- Role-based access control
- Export capabilities
Tech Stack: Python, Plotly, PostgreSQL, FastAPI
🎓 Text Classification with Hybrid CNN-GRU (Academic Project)
[GitHub Repository]
Research project on Indonesian text classification using hybrid deep learning
approach combining Convolutional Neural Networks and Gated Recurrent Units.
Tech Stack: TensorFlow, Python, NLP
Associated with: Universitas Airlangga
Get in Touch
📧 Email: fiqih.fathor.rachim@gmail.com
💼 LinkedIn: linkedin.com/in/fiqih-fathor-rachim
🐙 GitHub: github.com/fiqihfathor
| Open to: Senior AI Engineer |
ML Engineer |
|
| Location: Jakarta, Indonesia |
Remote |
Hybrid |