About
I’m Aditya Sampath, a software engineer with expertise in building scalable machine learning
and generative AI applications.
Just Completed my Master’s in CS from UNC Charlotte, specializing in deep learning, generative
AI, and LLM applications.
As part of my thesis, I'm evaluated the efficiency of building scalable systems using
LLMs for Computational Social Science, developing scalable annotation pipelines for high-risk
posts and language agents combined with temporal knowledge graphs for socio-political event prediction.
Projects
Experience
GenAI Researcher
University of North Carolina at Charlotte
Jan 2025 - May 2025
- Conducted a thesis on evaluating LLMs to develop scalable systems for Computational Social Science, engineering a scalable Chain-of-Thought annotation pipeline to classify 3.5M+ high-risk posts (~80% manual acceptance accuracy) and developing agentic workflows with Temporal Knowledge Graphs for bias-aware, interpretable socio-political event prediction.
Software Engineer Intern - GenAI Applications
Honeywell
Dec 2024 - May 2025
- Engineered a modular agentic workflow using the Reflexion technique, increasing ticket classification accuracy by 13% for ambiguous support emails.
- Designed a dual-agent prototype integrating ReAct and Hidden CoT strategies to generate fast and detailed answers in Honeywell’s internal GenAI tool
- Improved document retrieval precision by developing advanced prompt structures using Chain-of-Thought (CoT) and Hidden CoT prompting for query rewriting tasks.
Machine Learning Researcher
University of North Carolina at Charlotte
Feb 2024 - Jun 2024
- Deployed text-to-pose estimation models using Docker and HuggingFace, enhancing human motion capture and accuracy by 30%.
- Developed scripts to generate 3D human mesh structures, creating a pipeline for data augmentation and generation.