Building production AI that solves real problems at scale
I'm an AI/ML Engineer with 6+ years of experience, including 3+ years building and shipping production-grade agentic AI systems for enterprise and financial services environments. My focus is always the same: fast, reliable, high-impact AI in production — not prototypes.
- 🤖 Specialist in: Multi-agent orchestration, LLM engineering, Voice AI, RAG pipelines, LLM fine-tuning
- 🏦 Currently: AI/ML Engineer at MBL Hightech — building 24/7 AI systems for regulated financial brokers
- 📦 Open Source Author: DeepOCR — Python library for production document OCR
- 📄 Springer Published Researcher — IoT & AI systems (DOI: 10.1007/978-3-030-99581-2_3)
- 🎓 MSc Artificial Intelligence Engineering — Near East University
| Domain | What |
|---|---|
| 🧠 Agentic AI | Multi-agent systems with LangGraph, tool-calling, memory, conditional routing |
| 🎙️ Voice AI | STT + LLM + TTS pipelines, telephony integration, real-time call analytics |
| 📚 RAG Systems | Multi-stage semantic retrieval, vector stores, document intelligence |
| ⚙️ MLOps | CI/CD for ML, canary deployments, drift detection, automated retraining |
| 🔬 LLM Fine-tuning | LoRA/PEFT, training pipelines, SageMaker deployment, quantisation |
| 📊 LLM Evaluation | Benchmarking frameworks, hallucination detection, retrieval quality scoring |
| 📄 Document AI | OCR pipelines, KYC/ID verification, entity extraction, NLP |
- ✅ Deployed 24/7 AI customer support for a regulated financial broker — thousands of interactions daily, zero human agents, zero downtime
- ✅ Built Speech-to-Text analytics platform processing 1,000+ hours of audio/day at 95% transcription accuracy
- ✅ Achieved 60% reduction in RAG query latency and 60% reduction in client operational workload through AI automation
- ✅ Published open-source DeepOCR on PyPI — production-grade document OCR for developers
- ✅ Springer peer-reviewed publication — presented at international academic conference
AI & LLMs
Agentic Frameworks
Cloud & MLOps
Data & Vector
APIs & Integrations
Monotorinf
| Project | Description | Stack |
|---|---|---|
| 🔬 DeepOCR | Published Python OCR library on PyPI | Python, DeepSeek OCR |
| Phonenix AI | Enterprise-grade AI voice automation platform for businesses + Livekit and Twilio Intergration | |
| Whisper STT inference | Built with Whisper and Pyannote.audio for multi-language transcription and speaker diarization. + Runpod | |
| Ollama Inference Engine | Ollama API server + LLM + Embedding model | |
| 📶 Wi-Fi IoT Attendance System | Springer-published IoT research project | IoT, Cloud, Python |
Wi-Fi Attendance System in the IoT Era Springer · Forthcoming Networks and Sustainability in the IoT Era, pp. 19–29 · January 2022 DOI: 10.1007/978-3-030-99581-2_3
DeepOCR — A Python library for high-accuracy document OCR powered by the DeepSeek OCR model. Built for production document processing pipelines: scanned documents, IDs, forms, and financial records. Roadmap: hosted OCR API service for developer integrations.
pip install deep-ocr
Open to collaboration, consulting, and interesting problems in production AI.