Stealth AI startup
Work on backend systems, agentic evals, mcp, custom harness, post training
backend · post training · Agents · MCP · math
A record of the systems I’ve worked on and the engineering problems behind them.
Work on backend systems, agentic evals, mcp, custom harness, post training
backend · post training · Agents · MCP · math
At Johns Hopkins, my work spans GPU-accelerated XPBD soft-body simulation using NVIDIA Warp, real-time skin deformation, and differentiable physics calibrated with real-world displacement information. Graph neural networks form another part of this surgical-robotics research. Alongside physical simulation, I work on AI agents, agentic retrieval, MCP, and structured grounding.
Differentiable physics · Surgical robotics · GPU simulation · Graph neural networks · AI agents · Agentic RAG · MCP
At Synergetics AI, I worked on production MCP servers and the infrastructure around AI agents. The engineering work included client integrations, backend performance, and cloud infrastructure across GCP and AWS, with Terraform for infrastructure configuration.
MCP · Agent infrastructure · GCP · AWS · Terraform · Backend systems
At Reliance Jio, I worked with React and Spring Boot on internal operational software. API integrations, notifications, and workflow software connected the interface to the systems behind it.
React · Spring Boot · API integrations · Workflow software
My research internship at the University of Auckland focused on natural-language processing: extracting and classifying entities with CLAVIN and Stanford NER.
NLP · Entity extraction · Entity classification · CLAVIN · Stanford NER
AI agents · Agent evaluation · Model evaluation · LLM evaluation · Post-training · Backend systems · Production AI systems · AI reliability · Agent infrastructure · Tool-using agents · Model pipelines · Structured retrieval · Observability · MCP · LLM infrastructure
PyTorch · Differentiable simulation · Graph neural networks
Python · TypeScript · C++ · Java · Next.js · Node.js · Spring Boot · PostgreSQL · Docker · Terraform · GCP · AWS · Azure · Unix/Linux