Machine Learning Engineer  ·  Generative AI · RAG · Multi-Agent Systems

I turn business problems into production AI systems:- agentic pipelines, retrieval platforms, and ML services that ship and hold up under real workloads.

Machine Learning Engineer with ~4 years of experience delivering production-grade AI across GenAI, RAG, NLP, computer vision, and classical ML. I currently own an international client engagement, building autonomous multi-agent systems on Google Cloud. My focus is the unglamorous part that actually matters: reliability, grounding, evaluation, and getting models past a prototype into something a business depends on.

• 🔭 Now: Architecting production multi-agent systems (Google ADK, Gemini 2.5 Pro) on GKE
• 🧠 Depth in: RAG architecture, LLM systems design, agent orchestration, MLOps
• 🌍 Open to: AI/ML Engineer roles with visa sponsorship (Australia, EU, UK)
• ✍️ Building publicly: deployable, real-world tools not tutorial clones

Faunari — Venomous Snake Identifier PyTorch BioCLIP ONNX Streamlit GitHub Actions A recall-first, safety-critical classifier reaching 99.8% recall on venomous species (98.4% on a hard look-alike set) via cost-sensitive fine-tuning and isotonic calibration. Mahalanobis-distance OOD gate and a "never-say-safe" decision threshold; a CI/CD release gate blocks any model that regresses below the recall bar.

DocuSage — Multi-Tenant RAG Platform FastAPI React/TS Qdrant PostgreSQL Redis Celery Production-grade, multi-tenant RAG serving source-cited answers across 6 file formats. Hybrid retrieval (Reciprocal Rank Fusion) + cross-encoder reranking + a confidence-based abstention gate to prevent hallucination. Multi-agent orchestration (router → researcher → writer → summarizer → validator) with a programmatic citation validator that enforces grounding at zero extra LLM cost.

Languages   Python · SQL

GenAI / LLM   RAG · AI Agents · Multi-Agent Systems · Prompt Engineering · Fine-tuning (LoRA/QLoRA/PEFT) · RLHF/DPO · MCP

Frameworks   LangChain · LangGraph · LlamaIndex · CrewAI · AutoGen · Hugging Face Transformers

Retrieval / Vector   FAISS · Pinecone · Qdrant · ChromaDB · Weaviate · pgvector · Hybrid Search · Reranking

Eval / Observability   RAGAS · DeepEval · Splunk

ML / DL   PyTorch · TensorFlow · Scikit-learn · XGBoost · LightGBM · Computer Vision · NLP

Cloud / MLOps   GCP Vertex AI · AWS (Bedrock, Textract, S3) · Azure OpenAI · Docker · Kubernetes · Git

Merged pull requests into Hugging Face Transformers (#45752) and scikit-learn (#33853).

🥇 Vonage Hackathon 2023 (1st)  ·  🏅 AWS GenAI Hackathon 2024 (4th)  ·  🎖️ AWS GenAI Bonanza 2025 (Special Mention)

Reliability over demos. Grounding over guesses.