Senior AI/ML Engineer · 6+ years building production GenAI platforms, LLM systems and the backend and data infrastructure underneath them.
I design LLM-powered applications that hold up in production: multi-provider routing with fallbacks, safety guardrails, evaluation you can trust, and cost and latency budgets that survive real traffic.
- 🔭 Currently building enterprise-scale GenAI and LLM systems on GPT and Azure OpenAI, focused on reliability, latency and cost
- 🎓 MS in Information Technology, University of North Carolina at Greensboro · AWS Certified Solutions Architect
- 💼 Open to roles in AI/ML Engineering · Generative AI · LLM Systems · Backend & Platform Engineering
| Project | What it does | Stack |
|---|---|---|
| llm-gateway | OpenAI-compatible proxy that routes across OpenAI, Anthropic and Groq, tracks cost per team, enforces rate limits and serves repeats from a semantic cache — with no client code changes | Python · FastAPI · Redis · Prometheus · Docker |
| rag-platform | Multi-tenant RAG with hybrid retrieval (dense + BM25 + RRF), per-org isolation, citations and confidence scores, LLM-as-judge evals and hallucination detection | Python · FastAPI · Qdrant · Postgres · Langfuse |
| agent-framework | Agentic workflow engine: ReAct and planner agents, sequential / parallel / hierarchical / router orchestration, three-tier memory, dynamic tool registry | Python · FastAPI · OpenAI |
| contract-review-agent | Fault-tolerant multi-agent pipeline for contract review: stage checkpointing and resume, confidence-gated human review, full audit trail | Python · FastAPI · SQLAlchemy · Postgres · OpenAI |
| ml-platform | Full ML lifecycle: experiment tracking, LoRA/QLoRA fine-tuning, feature store with point-in-time joins, model registry, drift detection (PSI / KL / KS), LLM safety evals | Python · FastAPI · OpenAI |
| loop-pulse | Engineering insights from git history: truck factor and knowledge risk, contributors, activity trends, plus an LLM narrative whose every number is checked against the data | TypeScript · Fastify · SQLite · React · Docker |
All projects are tracked on the AI/ML Engineering Portfolio project board.
GenAI & LLMs — prompt engineering, model routing and fallbacks, RAG (hybrid search, re-ranking), agents and orchestration, LoRA/QLoRA fine-tuning, evaluation and safety, LangChain, Azure OpenAI
Backend — Python, Java, FastAPI, Spring Boot, REST and GraphQL APIs, microservices, event-driven systems
Data — Spark, Airflow, AWS Glue, SQL, real-time pipelines, vector databases (Qdrant, Chroma)
Cloud & ops — AWS (Lambda, Step Functions, ECS, SQS, EventBridge), GCP, Docker, Kubernetes, Terraform, CI/CD, Prometheus, MLflow
- 💸 Reduced LLM operational costs by 30% through optimization strategies
- 🛡️ Improved system reliability and cut MTTR by 40%
- ⚡ Reduced data latency by 35%, increased ingestion throughput by 45%, and improved pipeline and query efficiency by 30%+
📫 Connect with me on LinkedIn (link in the sidebar), or open an issue on any project.