May 2019 – Aug 2019 · 3 mos

• I worked as a GSoC student developer in 2019 for OpenCV (open source computer vision) platform under the guidance of Satya Mallick, Interim CEO, OpenCV and project details are mentioned below :
• 1. Implemented python bindings for facial landmark API in OpenCV comprising of models based on Active Appearance Model (AAM), Local Binary Features (LBF), and Ensemble of Regression Trees.
• 2. Explored the possibility of adding a 3D facial alignment model (based on an ICCV 2017 paper) to OpenCV Model Zoo.
• 3. Worked on making smaller facial landmark detection model with 5 points instead of 68 points.

May 2019 – Jul 2019 · 2 mos · Bangalore, India

• 1. Analysed the digital signage video to give insights regarding the gender, age, emotion of the audience and send them personalized recommendations.
• 2. Integrated facial recognition [based on the FaceNet paper] to identify face id and performed facial clustering using DBScan to add new face ids to the database and used this information to estimate the reach of the advertisements.
• 3. Implemented gaze tracking [based on Deep Pictorial Gaze Estimation paper] to estimate the average attention time of the audience on the digital signage board.

Mar 2019 – Mar 2020 · 1 yr

• Lead a team of 24 mentors to help them guide CSE students under the DAMP program and maintain regular interaction between the faculty and the mentors to smooth the mentoring process.

Apr 2018 – Apr 2019 · 1 yr

• Responsible to guide a group of 6 sophomores in academics and help them cope with their curriculum.

Apr 2017 – Apr 2019 · 2 yrs

• Represented the IIT Bombay CSE Batch of 2020 in various department-specific meetings and acted as a point of contact between the faculty and the students.

May 2018 – Jul 2018 · 2 mos · Pune, India

• 1. Developed a messenger racing game in Typescript using Pixi JS engine &Phaser framework.
• 2. Implemented backend server of game on AWS Lambda & used Webpack to bundle all assets.
• 3. Implemented localization & generated track from random seed using Perlin Noise Algorithm.
• 4. Developed messenger bot using webhooks & Graph API and deployed it on Heroku server.
• 5. Implemented trackers for data analysis & created UI, triggers & score manager for leaderboard.
• 6. Added race bots in game & simulated their motion based on player's skills and difficulty level.

Nov 2017 – Dec 2017 · 1 mo · Bangalore, India

• 1. Developed a prototype for calculating the stone deduction from jewels’ images using computer vision and image processing, helping in the calculation of the net weight of gold from the gross weight of the jewel.
• 2. Used background subtraction algorithms like Rolling Ball Algorithm for efficient detection of background, employed clustering algorithms like K-mean and DBScan for noise removal in result.
• 3. Made the prototype more robust by choosing multiple backgrounds depending on color of stones in jewel and finally mined existing data to convert found area into total weight of stone.
• 4. Implemented a testing framework for the application and improved the overall build and infrastructure of the prototype, achieving an accuracy up to 80%
• 5. Used AWS S3 for raw image storage and deployed the prototype as a web API using Node.JS framework on an AWS EC2 instance.

Jul 2017 – Jul 2020 · 3 yrs · Mumbai, India

• Assisted the professor in guiding 140+ students, creating and evaluating the exams, and conducting tutorials and extra classes.
• Teaching Assistant for Database and Information Systems (Theory + Lab) in Spring 2020.
• Teaching Assistant for Software Systems Lab in Autumn 2019. (Received best TA Award)
• Teaching Assistant for Software Systems Lab in Autumn 2018. (Received best TA Award)
• Teaching Assistant for Quantum Physics in Autumn 2017.