cv

Research Intern

Perception & Intelligence Lab, EE, IITK

Explored unsupervised & supervised learning for Radar Signal Deinterleaving in dense multi-radar environments.

• Built a MATLAB-based simulation GUI to generate labeled radar datasets under realistic tactical formations.
• Benchmarked HDBSCAN, Laplacian Graph-based Clustering, YOLO11n, YOLO11s on a custom multi-emitter dataset.
• Achieved 0.85 mAP with YOLO11 on real signals; tested novel stacked I/Q input for improved performance under signal overlap.

Freelance Machine Learning Engineer

Giovanni Franzan

Designed novel ML pipelines and algorithms.

• Benchmarked MediaPipe, ONNX.js, tfjs for low-latency, browser-based palm detection in a Unity app.
• Designed a novel 16-class Facial Emotion Recognition (FER) pipeline using MediaPipe Face landmark vectors → lightweight CNN, optimized for low-data training.
• Achieved 70% top-1 accuracy with real-time performance on a skewed 1K-image sample from a Google Research dataset.

Freelance Deep Learning Engineer

Fiverr

Delivered freelance projects focused on Web AI using TensorFlow.js, MediaPipe.

• Participated in a private beta of Neural Magic's model pruning and sparsification software.

Team Head — Avionics & Payload

IITK Rocketry and Space Exploration Team (RaSET)

Manages a 3-tier team of 50+ students alongside 3 other heads.

• Led development of two iterations of STM32-based SMD Flight Computer PCBs with onboard altimeter, IMU, GNSS, LoRa, and pyrotechnic ignition.
• Raised INR 7 Lakhs in sponsorship from IITK alumni.
• 2nd place internationally at Teknofest Rocket Competition (Turkey) in the Critical Design stage; won $2000 financial award.
• Cleared Critical Design Review (CDR) for Teknofest (Turkey) & IN-SPACe Model Rocketry.
• Developed avionics for 1 km apogee rockets for IN-SPACe and CANSAT competitions.
• Passed all flight simulations during Launch Readiness Review.

Secretary

Programming Club, IIT Kanpur

Led projects including Scam Detection under UP Police and an LLM Hosting project.

CBSE XII

NPS Yelahanka, Bangalore

Science

B.Tech

IIT Kanpur

Electrical Engineering

• Analog Electronics (A*)
• Ordinary & Partial Diff. Equations (A*/A*)
• Control Systems (A*)
• Introduction to Electronics (A*)
• Intro to Machine Learning (A*)

CBSE X

NPS Yelahanka, Bangalore

General

Academic Excellence Award

IIT Kanpur

For the 2023–24 session at IITK

Sri Singhasan Singh Scholarship recipient

Shri O.P. Jindal Engineering and Management Scholarship (OPJEMS)

2nd place at Freshers Case Study Competition

Combating Catastrophic Forgetting

Developed a novel Weighted Prototype Update for a CIFAR-10 classifier (on BeiT features) that generalizes to same-distribution datasets without labels.

• Prevented catastrophic forgetting, limiting worst-case accuracy to 97.4%
• Designed a Clustering-based Mean Shift prototype update algorithm for Unsupervised Domain Adaptation
• Achieved ~5% accuracy improvement on out-of-distribution data while maintaining performance on prior data

GenAI User Interaction Extension

Built a Chrome Extension tracking 12+ user interactions and analyzing attention proportions via LLM.

• Integrated webGazer.js for Eye Gaze tracking with automatic calibration through the device webcam
• Designed a pipeline to pass multimodal interaction data + HTML elements to Gemini and display results in console

Detection of AI-Generated Images

Inter IIT Tech Meet 13.0 - Classify AI-generated images, identify artifacts, generate explanations.

• Implemented a BNN with Fourier Magnitude and Local Binary Pattern channels for classification
• Benchmarked Static & Dynamic Quantization (PTQ & QAT) in PyTorch for real-time performance
• Used Jina-Clip-V2 and ResNets for artifact identification; fine-tuned Ovis1.6 for explanation generation
• Results: 2.6× speedup (80ms → 30ms) with <1% accuracy drop using Dynamic FX Mode PTQ; 97.2% accuracy, 30ms CPU inference