I'm an AI/ML Engineer who loves building practical AI systems — the kind that don't just sit on papers, but actually work in production.
By day, I’m architecting scalable LLM + RAG + Agentic AI pipelines for enterprise use cases at UNIS (3PL / Logistics AI).
By night (and sometimes weekends), I’m tinkering with GraphRAG, Neo4j-powered tool graphs, multimodal AI, and the occasional diffusion model.
Oh, and I’m also passionate about 🧮 math, 🐈 cats, and 🤖 agentic AI.
- LLM Engineering: Fine-tuning, PEFT (LoRA, QLoRA), quantization, model distillation
- RAG Architectures: Self-RAG, GraphRAG, RAPTOR-based retrieval, Neo4j + VectorDBs
- Agentic AI: MCP server orchestration, LangGraph-based tool agents, dynamic tool retrieval
- MLOps & Deployment: MLflow, Docker, Kubernetes, AWS SageMaker & Bedrock
- Data Science: NLP, CV, RLHF, Speech Processing, Recommender Systems
-
Unified MCP Tool Graph
GraphRAG + Neo4j + LangGraph for smarter agent tool selection. -
MedGPT
Multimodal LLaVA-based medical vision + text QA with HIPAA/GDPR compliance. -
Diffusion-based TTS with D3PM
SOTA speech synthesis + voice cloning with faster inference.
-
AI System Specialist - UNIS (2025-Present)
Building AI-driven solutions for logistics & supply chain optimization. -
Research Engineer & TA - CSULB (2023-2025)
Optimizing LLMs and teaching GenAI. Also wrote a thesis on Diffusion-based TTS. -
AI/ML Software Engineer - Cognizant (2021-2022)
Developed pilot scheduling optimization systems for Alaska Airlines.
- I’m an AI engineer who still loves pure math — topology and number theory fascinate me.
- I’ve trained agents to pick tools smarter than some humans (I kid… or do I?).
- I believe Agentic AI will change how we build software — one tool call at a time.
- I have an unreasonable number of cat memes in my downloads folder.
“The best AI systems feel like magic but work like math.” — probably me, during a late-night coding session.