I work on cloud infrastructure and AI systems, and I keep a strong hobby focus on astronomy and physics.
ML engineer focused on applied AI, agents, and robotics.
I am an ML engineer with interests in RL, agents, MLOps, and LLMOps, plus hands-on robotics projects.
Highlights
Achieved >85% task completion in 3D Unity/Gym environments using a custom ReAct agent. Optimized perception-to-action loop with LangChain.
Integrated Deep Learning perception models directly with ROS nodes for physical actuation. Solved Inverse Kinematics synchronization for a custom wheeled platform.
Achieved sub-100ms P50 latency for knowledge retrieval using LangChain + Pinecone + FastAPI.
View DetailsDeployed PyTorch inference on Azure GPU with autoscaling, cutting vendor API costs by ~60%.
View DetailsArchitected an event-driven compound AI system using Microsoft AutoGen and LangGraph.
View DetailsImplemented GRPO (Group Relative Policy Optimization) to fine-tune Gemma-3-270M for Chain-of-Thought reasoning.
View DetailsImplemented BridgeAttention architecture to map multimodal inputs (vision + language + proprioception) to 43-D action spaces.
View DetailsDeployed Octo-1.5 robotics foundation model on Hugging Face Spaces.
View DetailsEngineered a specialized VLA dataset that has heavily impacted the community (1,500+ downloads in 1 month).
View DetailsBuilt an autonomous agent navigating 3D Unity/Gym environments with >85% task completion rate.
View DetailsArchitected a production-ready backend handling WebSocket streaming with <150ms audio round-trip latency.
View DetailsIntegrated Deep Learning perception models with ROS nodes for physical actuation.
View DetailsEngineered a 4x super-resolution pipeline (SR-GAN) for hyperspectral satellite imagery.
View DetailsImplemented decentralized control policies for multi-intersection traffic management.
View DetailsBuilt a custom offline C++/Python engine for calculating Forward/Inverse Kinematics.
View DetailsImplemented and visualized 3D Forward Kinematics transformations from scratch.
View DetailsRetrieval Augmented Generation (RAG) lets you enrich LLM prompts with your own knowledge base. You embed documents into vectors and search them at runtime to pull in context. The question is: do you really need an external vector database? If you're just experimenting or working with small datasets, keeping everything in memory might be all you need.
ReadWelcome to my AI Engineering blog! This space is dedicated to sharing practical insights, experiences, and deep technical knowledge about AI infrastructure, cloud architecture, and modern development tools.
ReadAn overview of practical patterns for implementing Retrieval Augmented Generation (RAG) in production systems.
ReadExploring networking patterns for deploying large-scale AI workloads in Azure.
ReadOpen to work and collaborations. Reach me on GitHub, Hugging Face, or LinkedIn.