Building intelligent systems for complex environments.

Undergraduate Researcher: LLM Evaluation & Benchmarking

• Curated and annotated high-quality, multi-domain benchmark datasets across diverse categories, including Medical and Religious domains, ensuring robust data preparation for model ingestion.
• Conducted comprehensive evaluations of state-of-the-art generative language models (e.g., LLaMA, Qwen, Mistral) to systematically track response accuracy and answer correctness.
• Analyzed model performance limitations and behavioral edge cases, focusing on measuring hallucination rates and evaluating domain-specific knowledge alignment.

Undergraduate Researcher: Autonomous Driving & Simulation

• Tested and benchmarked diverse open-source autonomous vehicle control and end-to-end planning models within the CARLA simulation framework.
• Integrated system nodes with Autoware and ROS2 stacks to evaluate vehicle behaviors, lane-keeping, and routing under various environmental conditions.

Undergraduate Researcher: Autonomous Driving & LLM Agents

• Developed agentic decision-making frameworks based on Generative LLMs for autonomous vehicles as an alternative to traditional rule-based behavioral planners.
• Designed integration pipelines connecting high-level LLM agent outputs with ROS2 / Autoware lower-level planning and control modules.
• Reviewed contemporary literature on foundation models for robotics to draft simulated scenarios and evaluation suites for testing planning logic.

Undergraduate Researcher: NLP & Large Language Models

• Conflict Prediction: Utilized transformer-based architectures (BERT, GPT-style models) to parse and process political speeches and social-media texts for predicting underlying tension patterns.
• Cross-platform Profile Mapping: Implemented feature extraction and representation learning workflows to map user profiles across platforms using textual alignment.
• Big Data Analysis: Explored distributed processing architectures and engineered medium-scale preprocessing pipelines to handle textual data challenges.

Frontend Developer

• Built responsive web applications using React and Next.js (SSR/SSG) with Tailwind CSS, adhering to modern ES6+ and clean code standards.
• Integrated GraphQL queries/mutations and RESTful APIs, implementing proper caching, pagination, and accessible RTL/LTR component layouts.