Naveenraj Kamalakannan

I am a Master's student in Computer Engineering at New York University. This summer, I am an AI and Data Science Associate Intern at J.P. Morgan, where I build agentic applications to mimic human behavior for Asset and Wealth Management lines of business. Currently, I am researching advanced video analysis methods for stroke rehabilitation assessment, focusing on sub-second temporal resolution action detection at NYU Center for Data Science and NYU Langone under the supervision of Prof. Carlos Fernandez-Granda and Victor Li.

Research Interests:

• LLM Optimization and Alignment
• Reasoning and Planning
• KV Cache Optimization
• Efficient and Distributed Training
• GPU Acceleration Frameworks
• Reinforcement Learning
• Multi-Modal Retrieval Systems

I earned my Bachelor's in Electronics Engineering from Vellore Institute of Technology, where I collaborated with Prof. Sudhakar MS on medical image processing, developing the Exponential Pixelating Integral Transform for chest X-ray abnormality detection.

Email  /  GitHub  /  Google Scholar  /  LinkedIn  /  CV

Research & Publications

I'm interested in distributed training, model optimization, reinforcement learning and AI applications in finance, healthcare and robotics.

Exponential Pixelating Integral Transform with Dual Fractal Features for Enhanced Chest X-Ray Abnormality Detection

Built a chest X-ray abnormality detection model using Exponential Pixelating Integral transforms and fractal features. Implemented Multivariate Adaptive Regression Splines (MARS) ensemble, achieving 99.63% accuracy and F1 scores up to 98.10%.

A Novel Approach for the Early Detection of Parkinson's Disease Using EEG Signal

Led a team to develop a Deep Learning model analyzing EEG signals, achieving 93.3% accuracy in detecting early-stage Parkinson’s disease. Attained an F1 score of 93.48% and presented findings at the University of Tubingen Symposium.

Professional Experience

Real-world experience building AI systems, crunching data, and automating stuff across finance, industrial, and healthcare.

AI and Data Science Associate - J.P. Morgan

Architected and enhanced Asset and Wealth Management’s agentic platform, improving the relevance and naturalness of AI-generated content through a multi-stage retrieval pipeline. Orchestrated multiple AI agents with advanced reasoning capabilities to significantly enhance existing product functionality. Applied expertise in OpenSearch, RAG Systems, LangChain, KV Cache Optimization, and MCP Server to develop robust and performant AI solutions for financial applications.

Research Assistant - NYU Center for Data Science & NYU Langone

Working under the supervision of Prof. Carlos Fernandez-Granda and Victor Li on NSF-funded stroke rehabilitation research at the intersection of NYU Center for Data Science and NYU School of Medicine. Investigated and benchmarked state-of-the-art video understanding models (InternVL2, NVILA, LLaVa OneVision) for precise sub-second action detection in stroke rehabilitation scenarios. We came up with a multi-stage pipeline integrating YOLO11 for human-pose detection, DINOv2 Prompt-based Object Detection, and OpenMMLab’s pose estimation algorithms for motion analysis. Developed and integrated a Visual Language Model (VLM) based on video frames for improved contextual detection of patient interactions with objects, contributing to more effective and data-driven rehabilitation protocols.

Control Engineer - Zeeco Middle East

Built ML-based anomaly detection systems using Python, TensorFlow, and scikit-learn for industrial circuit analysis and equipment validation. Implemented predictive control algorithms for Pressure-Temperature-Flow (PTF) optimization using PID controllers and neural networks. Developed automated quality assurance pipelines with PyTorch and OpenCV for manufacturing process monitoring. Worked with PLC systems, SCADA, and industrial IoT protocols for real-time data collection and control.

Data Engineer Intern - Bajaj Finserv

Streamlined Azure-based data migration pipelines, reducing migration time by 34% and saving operational costs during the transition from EDW to Cosmos DB, leveraging Data Factory and Data Lake. Developed an ML model leading to ~33.3% reduced Data Warehouse Units resource consumption in Azure EDW.

Open Source Contributions

Contributions to major open source projects including Microsoft DeepSpeed and Snowflake ArcticInference.

FlashInfer Backend Support for SwiftKV - Snowflake ArcticInference

Added FlashInfer backend support for SwiftKV in PR #124, with automatic backend detection and improved throughput performance.

Bug Fix: Gradient Norm Calculation for CPU Offload - Microsoft DeepSpeed

Fixed a bug in PR #7302 where gradient clipping wasn’t working properly with CPU offloading in ZeRO-3. Added unit tests to cover different precision modes and gradient clipping scenarios.

Notable Projects

Cool projects that actually won things and made people go "wow, that's neat!"

Early Detection of Sepsis - National Hackathon Winner

Led a cross-functional team to develop a sepsis onset detection model using critical biomarkers (PCT and MDW), securing first place in the Design Category and winning a $2,000 grant from the VIT Incubator. The project focused on early detection of sepsis, a critical medical condition requiring rapid intervention.

Other Projects

Fun side projects where I get to play with robots, build AI stuff, and generally tinker with cool tech.

Starbots.AI Automated Cafeteria System

Built an autonomous mobile robot using ROS2, RRT* path planning, and OMPL for navigation. Integrated CNNs for object detection, SLAM for mapping, and robotic manipulation for food handling. Developed with Python, C++, and Gazebo simulation as part of The Construct Robotics Bootcamp.

Adaptive Monte-Carlo Localization Warehouse Robot

Built an autonomous warehouse robot using Adaptive Monte-Carlo Localization (AMCL) and Cartographer SLAM for precise positioning. Implemented Nav2 navigation stack with costmap generation, path planning, and obstacle avoidance. Developed with ROS2, Python, and Gazebo simulation for the RB1 robot platform.

Design and source code from Jon Barron's website