I am a Computer Engineering graduate building production AI systems where models meet real data engineering: retrieval pipelines, OCR-heavy document workflows, multi-model routing, and agentic products that need to be fast, observable, and usable.
|
AI systems RAG, vector search, OCR, local/cloud LLM routing, cited answers, and evaluation loops. |
Product engineering FastAPI, React/Next.js, TypeScript, Docker, PostgreSQL, MongoDB, Firebase, and CI. |
Automation LangGraph agents, Playwright/RPA, n8n workflows, market-data pipelines, and analytics. |
flowchart LR
A["Raw data<br/>docs, APIs, market feeds"] --> B["Pipelines<br/>OCR, ETL, WebSocket"]
B --> C["Retrieval<br/>MongoDB Vector, ChromaDB, FAISS"]
C --> D["Reasoning<br/>LLMs, LangGraph, multi-model router"]
D --> E["Products<br/>SAMETEI, NextHire, FinSenti, SMTbot"]
|
Async 24/7 crypto futures scalper on Bybit V5. WebSocket-native pipeline, Pine v6 emulator, Optuna tuning, and a private strategy core. ~24 ms cycle | 25 pairs | ~6700x speedup |
On-prem RAG assistant for enterprise HR procedures with LibreChat, MongoDB Vector Search, local Ollama models, and Qwen 2.5-VL OCR. 70% faster document handling | cited answers | air-gapped friendly |
|
Financial sentiment MLOps pipeline with FinBERT, LoRA/PEFT, MLflow, FastAPI, and Next.js delivery. 91.1% accuracy | 0.90 F1 |
LangGraph-powered job application agent: ATS scoring, skill-gap detection, personalized cover letters, and kanban tracking. Multi-model routing | FastAPI | Next.js | PostgreSQL |
