Naveenraj Kamalakannan

Benchmarked SOTA VLMs (InternVL3, NVILA, LLaVa-OneVision) and engineered a pose-refined prompting pipeline that integrates YOLOv11 pose tracks with VLM context to extract sub-second motion primitives, achieving ~67.75 Edit Score (ES) on structured upper-limb rehab tasks.

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%.

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.

Building agentic AI applications for investment workflows, taking multi-agent systems from prototype to scalable production deployments on the firm’s agentic platform. Driving performance engineering of the LLM serving stack: latency/throughput profiling, inference optimization, and capacity planning for production agentic workloads.

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.

Working under the supervision of Prof. Carlos Fernandez-Granda, Prof. Heidi Schambra and Victor Li. Studied whether current VLMs can recover clinically meaningful upper-limb motion primitives from rehab videos. Benchmarked SOTA VLMs (InternVL3, NVILA, LLaVa-OneVision) and engineered a pose-refined prompting pipeline that integrates YOLOv11 pose tracks with VLM context to extract sub-second motion primitives, achieving ~67.75 Edit Score (ES) on structured upper-limb rehab tasks and exposing systematic failure modes on subtle hand-object interactions.

ML-based anomaly detection for industrial circuit designs and predictive control (PID + neural networks) for pressure/temperature/flow optimization on plant equipment.

Streamlined Azure data-migration pipelines (EDW → Cosmos DB), cutting migration time by 34%, and built an ML model that reduced Azure EDW resource consumption by ~33%.

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.

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.

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.