Machine Learning Engineer

2024 — 2024

Palo Alto, California, United States

(acquired by Salesforce)

◦ Developed a system for producing call statistics and call outline using LiteLLM and OpenAI GPT-4o

◦ Optimized and improved the prompts for existing use cases using Prompt Engineering techniques

◦ Added ability to monitor and debug LLM models using LangFuse

2022 — 2024

Palo Alto, California, United States

Developed, optimized and implemented machine/deep learning models for computer vision

Developed a system for automatic improvement of segmentation models using PyTorch and Active Learning

Developed a system that generates synthetic data for segmentation models using Transformers and Generative Adversarial Networks

2021 — 2022

United States

VSCode extension for fixing errors in code: https://tensorbox.ai/

Built language models (transformers) in Pytorch

Developed low-level CUDA kernels (GPU programming) to speed up inference in C++ by 5x

Created backend in NodeJS and front-end in React

2021 — 2022

San Francisco Bay Area

Noise removal AI for online meetings (Mac app) and audio (SDK): https://utterly.app/

Built ML models in Tensorflow

Developed latency-optimized code for inference in C++

Created backend in NodeJS and front-end in React

2019 — 2019

San Francisco Bay Area

Built machine learning models for solving the long-tail problem in autonomous vehicles

Used Tensorflow to build the models

Developed a Gstreamer plugin in C++ for processing video frames in real-time

Startup School Online

Worcester Polytechnic Institute