More Projects > Research, Competitions, OSS, Personal

Research & Competitions (excluding published projects, see Papers):
1. [Fall '24] Inter-IIT Tech: Won gold at the Inter-IIT Tech Meet in 2024. Helped develop an optimized foundation model-based approach to open-world SLAM. Helped implement an explanation classifier for AI-generated images.
2. [AGV Fall '24] Game Theoretic Planner for an F1/tenth Scale Robot: Systems-level optimization (mainly focused on getting a good non-convex solver to work real-time our old NVIDIA Jetson) for a game-theoretic planner. Plan to enter the F1/tenth competition at ICRA 2025.
3. [CMU Summer '24] Amelia: Intent Prediction for Airport Surface Operations: Designed some fast map matching algorithms, developed a LLM-based procedural bias heuristic for Inductive Logic Programming based on an English description.
4. [IITKGP March '24] Goodness of Fit in Distributed Binary Detection: Conducted a literature search of recent results on sample complexity in communication-constrained binary detection. Derived some elementary corollaries for GoF when there are 2 inferrers.
5. [IITD Winter '23] Domain Adaptation for Breast Cancer Detection in Indian mammograms: Wrote some internal scripts and analyzed their outputs to help diagnose the poor transfer performance of SotA object detection algorithms when transferring to Indian mammogram data. Helped implement focal modulation and designed an appropriate masking strategy.
6. [AGV Fall '22] Machine Learning Reproducibility Challenge 2023: Lead two teams of juniors on MLRC 2023, reproduced results from the papers of CLRNet and the other thing. Conducted lit review. Helped design ablation study, edited report for CLRNet. Designed experiments on computational efficiency for the other one.
7. [AGV Summer '22] ECCV 2022 Visual Inductive Priors for Data-Efficient Deep Learning Challenge: Tried to piecemeal tweak some baselines to achieve data efficient object detection. The best model that could fit on our limited compute got us 10th place out of 63 teams.
Personal & Open-Source (recent only):
1. Automation services and scripts for secure RDP access to a shared desktop computer
2. A VLM finetuned on Minecraft Texture Packs
3. A trajectory generator for experiments in neurosymbolic learning

Continued in the next section. Please scroll.

More Projects > OSS, Personal, Coursework

Personal & Open-Source (select only):
1. A case study on IP grabbers
2. A survey of game theoretic approaches to differential privacy (sadly someone had already done the idea I had, but... fun read)
3. This website :)

Course Projects:
1. [CS60081 Fall '24] Usable Security and Privacy: TBD
2. [EC61205 Fall '24] Communication Networks: TBD
3. [CS60112 Spring '24] Information and System Security: CTF Challenge: Participated in a large number of CTFs ranging from simple PHP query injection attacks to convoluted use-after-free heap tricks.
4. [EC60012 Spring '24] Advanced Operating System Design: Elements of a Distributed System: Used system calls to re-implement inter process communication, a rudimentary file system and read/write/copy functions, user-level process migration (using CRIU) and Maekawa's Algorithm.
5. [EC61xxx Spring '24] Deep Learning: SVD-LoRA: Implemented LoRA for finetuning. Proposed a new initialization strategy using SVD.
6. [CS60077 Fall '23] Reinforcement Learning: Dynamic Gridworld: Implemented a few online RL algorithms and deep RL algorithms for navigating a gridworld in a particular manner (the right manner remaining unspecified to the agent-- only rewarded). Showed that it does not work thanks to the sparsity of valid transitions, and that it does when learning from an expert.

Continued in the next section. Please scroll.

More Projects > Coursework

Course Projects:
1. [EC60007 Fall '23] Computational Neuroscience: Neural Signal Processing in MATLAB: Conducted a wide variety of modelling and simulation experiments on real data, from the Hodgkin-Huxley and Morris-Lecar equations to extracting maximally informative dimensions.
2. [CS60050 Spring '23] Machine Learning: Various benchmark tasks:
  - Rice Variety Classificaiton using Naive Bayes Classifier
  - Heart Disease Detection using Support Vector Machines
  - K-means clustering vs. Single-Linkage Top-Down Agglomerative Clustering
3. [EC60004 Spring '23] Neuronal Coding of Sensory Information: Processing Cat Auditory System Signals in MATLAB: Analyzing discrimination in cat auditory nerve fiber responses to tones and speech
4. [EC39002 Spring '23] Embedded Systems Lab: Temperature Controller on an 8051: Wrote a temperature controller in Assembly for an 8051 which turns a fan on whenever a thermometer reading crosses a certain threshold and off vice versa.
5. [CS60092 Spring '22] Information Retrieval: I forgot: I honestly can't for the life of me remember what I did and I didn't upload it to GitHub.
6. [CS61603 Fall '21] Computational Foundations of Cyber-Physical Systems: Differential Privacy in a Smart Grid: Implemented a smart grid in MATLAB-Simulink, ran simulations, added Gaussian noise-based differential privacy and re-ran those experiments. (In retrospect, I don't remember why it was Gaussian. Laplace might have made more sense).
7. [DY17003 Fall '20] DIY Project: Gesture-Controlled Medical Robot: Used pretrained models from MediaPipe and developed a sign language to control a robot. Sent control signals over WiFi to an Raspberry Pi robot.