experience


┌───────┐
│       │
│       │
│       │
└───────┘

alan turing institute

dec 2024 - april 2025 | research associate | collaboration: university of birmingham & u.s. army research institute

  • developed a heterogeneous-agent reinforcement learning (harl) framework using human-proxy agents that simulate realistic human constraints and capabilities to improve ai-human collaboration in multi-agent systems.
  • designed a cooperative grid-world capture environment based on stag hunt game theory, where machine agents had full observability but couldn't detect target health, while human-proxy agents had limited vision but unique disease detection abilities.
  • conducted experiments across various environment configurations, varying disease probability and penalty severity to analyze cooperation patterns.
  • demonstrated that rl agents trained with human-proxy teammates achieved superior cross-environment performance, with teams trained under moderate risk conditions showing 30-40% higher collaboration rates.

riskopsai

jun 2023 - aug 2024 | ai ml intern

  • built a deep learning pipeline with tensorflow and pytorch using resnet-50 and transformer models for classification. used tensorrt to optimize inference, achieving 25% faster performance and 15% higher accuracy.
  • created distributed ml infrastructure with apache airflow and mlflow on aws gpu clusters using horovod, reducing training time by 20%. set up automated data pipelines for feature engineering.
  • developed a predictive analytics system using postgresql and bigquery with scikit-learn and xgboost on imbalanced data. optimized queries to improve decision-making efficiency by 30%.

srm institute of science & technology

oct 2022 - feb 2024 | researcher under dr. vaishnavi moorthy

  • developed an autonomous navigation system using ros2 by fusing lidar, rgb-d, and imu data through an extended kalman filter. this improved localization accuracy by 15%.
  • built a slam system using sac and trpo algorithms in pytorch to improve path planning with rrt* and a*. reduced navigation errors by 25% using dynamic obstacle avoidance.
  • created a real-time perception pipeline using opencv and pcl, integrating yolov7 for object detection. achieved 20ms latency and 95% detection accuracy in changing environments.