AI Engineer building production-oriented LLM, RAG, agentic AI, ML, and data systems.

I build end-to-end AI products from experimentation to deployment, with a focus on reliable LLM applications, retrieval systems, and data-driven decision tools.

Multi-agent insurance claims assistant combining LLMs, RAG, and NL2SQL to provide accurate, context-aware answers from business data.

Focus: Multi-agent AI • RAG • SQL generation • Customer support automation

AI-powered recommendation system for risk assessment using graph-based reasoning and semantic retrieval to support decision-making.

Focus: Graph AI • Recommendation systems • Knowledge graphs • Risk analytics

Financial sentiment analysis pipeline with model tracking, retraining workflows, monitoring, and API serving for production use.

Focus: NLP • MLOps • Fintech AI • Deployment

Machine learning system for detecting abnormal network traffic and suspicious activity patterns.

Focus: Anomaly detection • Cybersecurity • ML

I contribute to projects at the intersection of Python, AI, LLMs, retrieval systems, and data engineering.

• LlamaIndex — contributed to document update behavior and safer async reference handling.
• VaultRAG — contributed an embedding-dimension validation fix to avoid incompatible embeddings reaching vector storage.

I’m especially interested in open-source work around LLM applications, RAG, agent frameworks, and practical ML tooling.

Claim management multi-agent system

• Built a conversational multi-agent assistant for health insurance claim handling, enabling natural-language access to policy and reimbursement data.
• Developed a reliable SQL-generation layer for structured, secure querying over business data.
• Improved system responsiveness and observability through embedding caching and dynamic model routing.
• Integrated the solution in a web interface for end-user interaction and business use.

02/2025 – 08/2025

Fraud detection and anomaly monitoring

• Developed anomaly detection models for payment transactions to flag suspicious activity and reduce merchant fraud.
• Served the model through a FastAPI application in Docker for continuous transaction scoring.
• Prioritized alerts using SHAP-based interpretation and recall-focused business evaluation.

07/2024 – 08/2024

Document intelligence and risk prioritization agent

• Built a document-search agent for internal project-risk prioritization using semantic retrieval and knowledge-graph techniques.
• Compared traditional RAG and GraphRAG patterns to improve answer quality and reliability.

Engineering Degree in Computer Science

Specialization: Data Science
Graduated with very good mention

Building AI systems, contributing to open source, and learning by solving real problems.