Product lead for:

1) Nemo Microservices - Leading a cross-functional team of 8 PMs, 40 engineers, and 100 solution architects. A microservice platform enabling enterprises to build gen A.I applications, including: Data curation, model fine tuning, model evaluation, information retrieval, inference at scale, and guardrails for safe A.I.

https://developer.nvidia.com/blog/simplify-custom-generative-ai-development-with-nvidia-nemo-microservices/

2) NVIDIA NIM - Optimized inference service for deploying genAI models: https://www.nvidia.com/en-us/ai/

3) Audio Language Model API - Productinizing latest research from NVIDIA'S Advanced Deep Learning Research team. Building new speech-to-speech models for conversational AI

4) Nemo Customizer - software to automate LLM fine tuning on kubernetes using LoRA, SFT, RLHF, DPO, P tuning

5) Triton Inference Server and Dynamo: open source software to deploy AI models in production at scale. Grew customer adoption by 615% YoY, resulting in over 3 million downloads to date.

https://developer.nvidia.com/nvidia-triton-inference-server

6) BioNemo - shaping NVIDIA’s roadmap for AI models for digital biology and drug discovery (e.g. 3D protein structure modelling, virtual cell screening, molecule generation).

Built AI assistant to help primary care physicians with diagnosis and treatment selection of mental health conditions

Built AI agent that transcribes patient conversations, matches symptoms against DSM-5 diagnostic manual, and searches APA psychiatric guidelines for medication tailored to the patient profile.

Using AI models to automate underwriting for insurance companies, saving up $10M in costs

• Developed ensemble models which predicts an applicant's risks core, with a classification accuracy up to 90%

• Created a software prototype to digitize insurance application forms, reducing labour by 67%

• Responsible for hiring, team management, raised investment funding, and led negotiations for an acquisition by a leading Canadian life insurance firm

• One of 36 individuals selected from thousands of applications across Canada (~3% acceptance rate).

• Next 36 is a national program that accelerates the growth of Canada’s top innovators

• Taught by faculty from Rotman School of Management, Harvard Business School, and MIT Sloan School of Management

• Developed ROS architecture and system design for various ML systems: computer vision, traffic light detection, lane detection, localization, sensor fusion, and control systems.

• Placed in first place two years in a row for the SAE AutoDrive Competition - an international AV competition where top universities around the world compete to build a level 3 autonomous car.