Technical Product Owner

• Architected an LLM based BOM assistant platform and led the development with Python, FastAPI, and PostgreSQL,containerized services with Docker, deployed on AWS EKS, & integrated MongoDB for chat and session persistence.

• Researched on fine-tuning, prompt engineering, and state management with LLM to build a multi-agent system with Anthropic Claude that follows ReAct paradigm, using LangGraph framework enabling robust async workflows, modular tool APIs, and dynamic prompt chaining. Expanding the platform beyond POC to production workflows and user needs.

Software Engineer

• Responsible for standardizing the repairability of the cover glass and face id module in iPhones. Tasks include developing, verifying software bundles of the modules, curating requirements for the new products, and corresponding with overseas production teams.

• As a Station DRI, I track the main production software developments, implement those in our test framework, approve Bill of materials (BOM) for new programs, and perform trouble shooting and analysis at the station level and program level.

• Achieving an yield rate of 100% in year-on-year build.

Project - Metaverse

• Worked on Capture systems comprising of multiple cameras, Kinect sensors and Bellus Arc system to generate 3D avatars, as a part of Holograms team for Metaverse project.

• Managed software developers and multiple vendors to develop a platform to control different systems and congregate the data.

• Brought up 1.5 PB volume Dell Isilon server and orchestrated the pipeline to gather participants data, and transmit data to Gaia server with Isilon as a buffer.

Designed and developed innovative robotic system solutions to automate internal processes ranging from sample preparation to analysis. Some of the systems are designed from ground up, some are retrofitted as a separate module to an already existing system and some are a combination of custom designed modules, cameras and of the shelf systems like robotic arms.

Worked on Cosmetic Grading of phones using Intelligent cleaning and Camera systems.

• Generated a surgery tool coordinate dataset with DeepLabCut python vision package to develop video analysis model.

• Developed an end-to-end Video classication pipeline with CNN and LSTM to create a video surgery dataset.

• Developed a coordinate predictor with LSTM (RNN) using the dataset from DLC tool.