Mar 2023 – Nov 2023 · 8 mos · Bengaluru, Karnataka, India · Hybrid

Oct 2021 – Mar 2023 · 1 yr 5 mos · Bengaluru, Karnataka, India · Remote

• Part of IBM’s first monolithic repository initiative that is leading an organization level culture shift towards a unified and collaborative development pipeline.
• Improved and Presented the POC for AI Enabled User Sentiment Prediction and Analysis.
• Refined the content-checker (Pre-merge scanner for detecting secrets) by identifying and eliminating 36% of false positive scenarios.
• Effectuated the Log4j package version upgrade for all the projects in the monorepo for the highly sensitive ”Log4Shell” vulnerability within a narrow time-span.
• Enabled usage-metrics collection for one of IBM’s flagship products ”Cloud Pak for Data”

May 2020 – Aug 2020 · 3 mos · United States · Remote

• Generated synthetic multi-spectral satellite imagery using state of the art technology to ameliorate the need for ground-reference data which can be
• expensive and laborious to collect.
• Eliminated artifacts in generated synthetic data through the use of multi-scale gradients.
• Pronounced improvement in visual quality via pixel-wise feature vector normalization and equalized learning rate.
• Achieved near 50% reduction in memory requirement and latency through appropriate modification of the architecture with minimal loss in quality

Jul 2018 – Aug 2019 · 1 yr 1 mo · Noida Area, India

• I was lucky enough to get a project in applied ML! I proposed and implemented a Generative Adversarial approach to solve the color matching problem in collaboration with teammates in Noida and experienced researchers in the US.
• Generative Adversarial color matching : Traditional machine learning techniques for the color matching problem require a dataset annotated by experts which can be extremely expensive and time-consuming to generate. A smart reformulation of this problem allowed us to use easily available datasets to train a GAN solving the color matching problem.

May 2017 – Jul 2017 · 2 mos · Bengaluru Area, India

• Developed a system that rapidly generates marketing segments given the database of users.
• Highlights:
• 1. Best Presentation and Demo Award.
• 2. Patent filed.
• 4. Exhibited clever use of data structures and algorithms to achieve a 100 time speedup.
• 5. Deployed autoencoders, RBMs, Factorization machines and Tree Based Methods.
• 6. Created a web application using HTML, CSS, JS, Python for the award winning final demonstration.

May 2016 – Jul 2016 · 2 mos · Kharagpur Area, India

• Guide : Prof. Partha Pratim Das, Computer Science and Engineering, IIT Kharagpur
• Developed a dance tutoring system to compare the dance of novice to that of an expert and point out the mistakes.
• 1. Implemented Dynamic Time Warping Algorithm to compare videos of differing frame lengths.
• 2. Defined and implemented a similarity measure for poses in two frames.
• 3. Performed validity checking for frames using traceability, sudden jerk detection and conformity of detected skeleton with color.
• Languages and Technologies : Matlab, Python, C++, Microsoft Kinect.

Jul 2015 – Apr 2016 · 9 mos · Kharagpur, India

• Lead the team that developed a crowd management software for event Alakazam.
• Developed in part the SF-IDOL page and also the Merchandise page of the fest.
• Designed in part a new and dynamic sponsors page.(https://2019.springfest.in/sponsors/)
• Languages and Technologies : HTML, CSS, Python, Flask, PHP.

Jun 2015 – Jul 2015 · 1 mo · Mumbai Area, India

• One of theory graders of the International Physics Olympiad 2015!