Making AI reliable through testing, observability, and clean MLOps.
I'm an AI/ML engineer passionate about building production-grade AI systems that are not just intelligent, but reliable, testable, and observable. My work spans the entire lifecycle — from data validation and model training to deployment, monitoring, and continuous evaluation.
- 🧪 AI Quality & Testing — Building frameworks to test, evaluate, and validate AI agents and LLM applications.
- 🔍 MLOps & Observability — Designing end-to-end pipelines with monitoring, drift detection, and performance tracking.
- 🤖 RAG & Agentic Systems — Developing retrieval-augmented generation agents and multi-agent evaluation frameworks.
- ⚙️ Full-Stack AI — Creating APIs, CI/CD pipelines, and cloud-ready platforms that bridge research and production.
| Project | Description | Tech Stack |
|---|---|---|
| AI Quality Ops | End-to-end AI Quality Operations platform with 5 automated pillars (data, accuracy, security, performance, observability), QA-as-a-Service API, and CI/CD. | Python, FastAPI, Docker, Cloud |
| synapse | Autonomous AI agent with LangGraph, tools (web search, SQL, calculator), memory, and guardrails. | Python, LangGraph, LangChain |
| model-hub | Real-time multi-modal content moderation API (text + image) with async inference and explainability. | Python, FastAPI, PyTorch, Transformers |
| fraud-detection | Real-time fraud detection API with drift monitoring (FastAPI, Kafka, Evidently, Prometheus). | Python, FastAPI, Kafka, Evidently, Prometheus |
| medical-rag | Medical Q&A API using RAG and Hugging Face Inference (FastAPI, Chroma, Docker). | Python, FastAPI, Chroma, Hugging Face, Docker |
| ai-agent-validation | AI Agent Validation Framework — LLM-as-a-Judge with mock MCP, Python/pytest. Cross-checks correctness, sentiment & hallucinations. | Python, Pytest, LLM APIs |
| multi-agent-test-sandbox | Multi-agent test sandbox: Doer, Adversary & Judge agents. Evaluates AI responses for security, bias, and correctness under prompt injection attacks. | Python, Pytest, CI/CD |
| pact-contract-testing | Consumer-driven contract testing using the Pact pattern. | Python, Pact, Flask, Pytest |
I'm always open to discussing AI reliability, MLOps, or potential collaborations. Reach out!
"AI that works is AI that is tested, monitored, and continuously improved."