• Hi!
    I'm Nikhil

Who Am I?

I'm Nikhil Bhat, an engineer at Amazon with experience in computer vision, robotics, and software engineering. I love building high-impact software solutions at scale. Check out my resume here! :-)

Work Experience

Software Engineer at Amazon
August 2021 - Present

• Re-architected a legacy classification service to support multiple clients, increasing throughput to 10M requests/day
• Migrated critical services from AWS AppConfig to AWS Parameter Store, reducing infrastructure costs by 30%
• Designed a scalable event-driven compliance service to process customer data deletion requests (50k requests/day)
• Developed a Python tool to root cause and resolve common operational issues, reducing on-call load by 5 hrs/week
• Drove org-wide adoption of AWS best practices, reducing incoming security vulnerability tickets by 100% (10/month → 0)
• Mentored an intern project to regionalize an event publisher used in 10+ microservices, leading to a successful return offer
• Led a team of 4 engineers across 3 timezones to onboard new use cases, processing an additional 130k orders/month

Software Engineering Intern at Amazon
June 2020 - August 2020

• Saved over $100k a week by expanding a native AWS service used to identify products being sold at a loss
• Built an automated service leveraging AWS Lambdas to alert business owners on recently suspended products

Software Engineering Co-Op at Intel
January 2020 - June 2020

• Implemented computer vision and data processing algorithms for 3D Athlete Tracking at the Tokyo 2021 Olympic Games
• Designed an end-to-end static camera calibration application using OpenCV and Tkinter resulting in under 5cm of error
• Decreased runtime of the biomechanical analysis module by 20x through algorithm optimizations and data processing adjustments
• Implemented a modified inverse kinematics algorithm on the biomechanical analysis module, reducing start of race error by 70%
• Deployed a modular unit-testing framework for verifying camera calibration accuracy, and integrated it into the team’s CI/CD pipeline

Software Engineering Co-Op at Greensight
January 2019 - July 2019

• Developed an indoor, vision-based, positioning system for drone navigation in GPS-denied environments by using OpenCV and Python to localize position based on a series of ArUco Markers. This earned the company $30,000 in the Verizon 5G Robotics Challenge.
• Designed a self-correcting algorithm for heliostat motion, by using OpenCV to account for GPS errors
• Implemented SLAM navigation algorithms using a combination of 2D LIDAR and depth cameras for an autonomous ground robot
• Created and deployed a ROS network of 100 interconnected devices to track the sun and generate solar power

Data Analyst Intern at Barnes Aerospace
May 2018 - August 2018

• Reduced turbine center frame inspection times by 30% by developing a Python program to implement statistical process control
• Created an interactive GUI using Python linked to an external SQL database allowing for supervisors to view consolidated error reports within any timeframe and location.
• Programmed an application in VBA to dynamically schedule manufacturing operations, accounting for variable lead times for part ordering and tool design

Research Assistant at Northeastern Field Robotics Lab
Sepetember 2017 - April 2018

• Programmed a real time kinematic GPS driver using C++, and implemented it into the autonomous car ROS environment
• Furthered development in drone navigation techniques utilizing AprilTag image detection with ViSP and OpenCV

Robotics Ambassador at CT Science Center
June 2014 - August 2015

• Increased accessibility for disabled patrons of the Hartford Farmer’s Market by manufacturing an autonomous shopping robot utilizing Arduinos, GPS Shields and IR sensors to follow the customer
• Designed and built a robotics exhibit involving driver-controlled soccer robots playing against an Arduino-based autonomous goalie

Projects

Fancily

Machine Learning Fashion App

MoCap

OpenCV Based Indoor Drone Navigation System

Waste Disposal Robot

Robotic Arm to Dispose Waste After 'Reading' a Note from the User

Endless Trivia

Trivia Game that Uses the Wikipedia API to Automatically Generate Questions

Don't Touch My OJ

OpenCV + Raspberry Pi Based Fridge Security System

Sesquipedalian

Words with Friends Solver which uses OCR to read the board.

Friends' Websites

Daniel Castle EE @ Apple
danielscastle.com

Lucas Briggs EE @ Astranis
lucasbriggs.me

Shruti Rajagopalan UI/UX @ Amazon
shrutir.com