I'm a Staff Machine Learning Engineer with 8+ years of experience bridging cutting-edge research and production systems. I've worked in enterprise GenAI platforms, multi-agent architectures, paper-to-production model implementation, time-series and large-scale ML infrastructure across finance, insurance, retail, real estate, and technology. Working from small startup environments to global enterprise teams across all continents.
- 📄 GenAI Document Processing at Scale: Building intelligent extraction systems processing 1M+ pages/month in production.
- 🤖 Enterprise Multi-Agent Systems: Architecting agentic workflows (LangGraph, PydanticAI, Bedrock, MCPs) with Human-in-the-Loop (HITL) execution.
- 🛠️ Self-Hosted LLM Infrastructure: Custom Kubernetes, kServe, and vLLM setups for secure, one-click internal LLM deployments.
- 🔬 Paper-to-Production Engineering: Translating ML research papers into production systems—from scale forecasting architectures (N-BEATS) to custom deep learning frameworks built from scratch.
- 🔍 Custom Retrieval & RAG: Architecting high-precision OpenSearch retrieval pipelines and custom scoring models.
- 🗺️ Wayfinder - A hybrid of a self-generating knowledge graph and the LLM Wiki for coding memory management.
- 🪨 Sisyphus - A lightweight task manager for MacOS - "A task manager for a never ending workload"
- 📄 local-doc-parser – Local, privacy-first document parsing and structured LLM extraction.
- 💻 code-share – VS Code extension acting as Live Share for coding agents.
- 📐 siamese-confidence – Pure NumPy implementation of a Siamese Neural Network.
- 🧹 anti-slop-python – Static analysis and quality guardrails to strip out AI-generated code slop.
- 🏷️ title-your-pr & describe-your-pr – Automated GitHub Actions to write clear, context-aware PR titles and descriptions.
- 🏛️ FastAPI Layered API Template – Production-ready, modular architecture template for scalable Python microservices.
- 🗣️ KubeCon AI Day: Making Complex R Applications into Production using Argo Workflows
- 📝 Argo Blog: Making Complex R Applications into Production using Argo Workflows
- GenAI & Agents: AWS Bedrock, LangGraph, PydanticAI, Claude SDK, vLLM, kServe, Opik, HuggingFace
- Machine Learning & Data: PyTorch, Scikit-learn, XGBoost, Databricks, PySpark, OpenSearch, Pandas, Numba, StatsModels
- Backend & Infra: Python, FastAPI, Docker, Kubernetes, Helm, Argo Workflows, PostgreSQL, MongoDB, DynamoDB
- Cloud & DevOps: AWS (ECS, Lambda, EventBridge, S3), GCP, Azure, GitHub Actions, Terraform
Feel free to message me on LinkedIn
- ⚡ Fun fact: I'm a Master of Science in Structural Engineering, having my research based around modelling wind loads using Computational Fluid Dynamics, so I guess I can answer regarding best practices using Finite Elements Analysis and Finite Volumes as well.