Bachelor's degree

Evaluated large language model (LLM) responses for correctness, reasoning quality, factual accuracy, and instruction following. 
 
Designed challenging prompts and edge-case scenarios to assess model capabilities across technical and analytical tasks. 
 
Ranked multiple model outputs using structured evaluation rubrics and provided detailed feedback to improve response quality. 
 
Identified reasoning errors, hallucinations, logical inconsistencies, and code-generation failures to support model alignment and fine-tuning. 
 
Performed human-in-the-loop evaluation of AI-generated content, contributing high-quality annotations used for reinforcement learning and model improvement. 
 
Collaborated on projects involving prompt design, response evaluation, and quality assurance for frontier generative AI models.

LLM Evaluator

- Built and maintained backend for a practice/competitive coding platform serving 3,000-4,000 active users;       optimized REST APIs for concurrent traffic

- Built Data Battleground, a real-time WebSocket competition engine supporting 400-600 simultaneous users with low-latency matchmaking and live score sync

- Designed MongoDB/Mongoose schema with indexing strategy for fast search, sort, and pagination across thousands of records 

- Automated MySQL query grading, eliminating manual review 

- Wrote a Python automation pipeline parsing DOCX -> bulk DB load, cutting content upload time by 95%

- Deployed and managed Dockerized services on AWS EC2 with Nginx/Caddy reverse proxies

Software Development Engineer | Jan 2024 – Dec 2025 | Remote

I am computer science grad, I have 2 years of experience in backend development. Worked in an early stage startup

github

Know about Poras Shrivas's work experience, education, bootcamps, certifications, and much more. Poras Shrivas