I build distributed systems that serve
millions of customers. Shipped features that enabled 11,000+ family invitations in 3 days
and improved platform reliability across multiple regions.
IIT Bombay · JEE Advanced Rank 711 · Published Researcher
01 — About
I'm a Software Development Engineer at Amazon where I build backend services that power account management and family sharing experiences for customers worldwide. Within my first quarter of joining, I was awarded Rookie of the Quarter across the entire Identity Services organization.
My annual review noted that I "significantly exceed SDE I expectations through exceptional technical delivery and ownership" — having launched a major product expansion within 6 months of joining as a new hire.
I graduated from IIT Bombay with a Dual Degree (B.Tech + M.Tech) in Electrical Engineering, with dual minors in Computer Science and Data Science. I secured All India Rank 711 in JEE Advanced (out of 1.2 million students).
Previously: Quantitative Research at WorldQuant (part-time), cancer classification research at Tata Memorial Hospital, and an award-winning Amazon internship.
02 — Experience
Amazon Photos & 6-Member Family Expansion
Customer Experience Improvements
03 — Projects
From IIT Bombay coursework, research, and personal exploration
Swin-Transformer model for retinal disease detection surpassing State-of-the-Art (mAUC 0.761 vs SuperRetina's 0.755). Reverse knowledge distillation approach.
LSTM-based trading model handling high-frequency, intraday, and interday scenarios. Generated 6x profit using historical OHLCV data across multiple stocks.
Thompson Sampling, UCB, KL-UCB for multi-armed bandits. Value Iteration, Linear Programming, and Howard Policy Iteration for MDP planning. Solved a multi-level billiards game.
Spacy NER + BART model achieving 98.9% precision on CNN dataset for entity-aware text summarization with integrated classification.
Autoencoder network for image deblurring achieving PSNR of 30 dB on 24,000 images with only 75,651 parameters and 90.29% test accuracy.
Open-source contribution to Manifold — a C++ geometry library for topological robustness used in 3D printing and CAD applications.
Led development of a hostel app with Django + Flutter featuring biometric food tracking and RFID-based food waste analysis. Available on Google Play Store.
GaAs nanoelectric device behavior analysis on COMSOL multiphysics. IEEE-format publication on device characteristics.
04 — Publications
Ravi Kant Gupta, Mohit et al. Attention-based deep learning for lymphoma subtyping from histopathology slides — AUC improvement from 0.81 to 0.87.
Sunny Gupta, Mohit et al. Novel federated optimization combining Nesterov acceleration with Newton sketch for communication-efficient distributed ML.
04.5 — Blog
Automatically pulled from my Medium
Loading posts from Medium...
05 — Skills
06 — Recognition
From peer feedback and annual reviews at Amazon
"Significantly exceeds SDE I expectations through exceptional technical delivery and ownership. Swiftly delivered complex features within months of joining."
"Completed the Photos 6-Member Expansion project in 6 months as a new hire. Owned critical workstreams end-to-end — always at the forefront of delivering on time."
"Rapidly ramped up on Amazon's complex distributed systems as a fresh college graduate and contributed immediately to production services."
"Displays extraordinary Customer Obsession — went beyond requirements during the Photos expansion and CX Papercuts projects to improve the customer experience."