Skip to main content
Ekagra Gupta

Ekagra Gupta

IIT (ISM) B.Tech graduate with 1.5+ years of professional experience in LLM evaluation, AI agent testing, and OCR/document-AI QA. Open to remote roles.

Talent
BokaroMember since April 5, 2026
PythonJavaScriptLLM EvaluationAI Agent TestingBenchmarkingNode.jsDjangoDjango REST FrameworkZustandTanStack QueryREST APIsJWT AuthenticationRBACMongoDBPostgreSQLAirtable
·

Hire this person

Send a job offer directly to this candidate

PythonJavaScriptLLM EvaluationAI Agent TestingBenchmarking

About

IIT (ISM) B.Tech graduate with 1.5+ years of professional experience in LLM evaluation, AI agent testing, and OCR/document-AI QA at Mercor and Outlier AI. Proficient in Python, JavaScript/React, Django, and Node.js with a strong foundation in DSA (LeetCode top 4%, Codeforces Expert). Available immediately for remote roles.

Experience

Software Engineer (CAE) & OCR Specialist - Mercor - Remote

(2025-10)

  • Assessed AI coding agents on 200+ production-grade codebases — evaluating correctness, reasoning quality, refactoring, test generation, and execution behavior across web, API, script, and multi-file projects.
  • Led benchmarking and regression-style reviews: compared agent behavior across model versions, isolated failure modes (silent bugs, partial fixes, API drift), and delivered actionable performance metrics to model teams.
  • Designed evaluation rubrics and playbooks ensuring scoring consistency across ambiguous specs, edge cases, and acceptance criteria; tracked results in Airtable for stakeholder prioritization and model iteration.
  • Improved text extraction accuracy on noisy scans, complex tables, and multi-column layouts through error taxonomies, structured correction workflows, and iterative QA feedback cycles.

AI Trainer & Code Evaluator - Outlier AI - Remote

(2025-01 - 2025-05)

  • Built LLM training and evaluation tasks in Python and JavaScript/React: debugging, feature implementation, refactoring, performance tuning, and code readability improvements across diverse problem sets.
  • Scored model outputs with multi-criteria rubrics (correctness, efficiency, style, security patterns); refined prompts using structured failure analysis to improve model reasoning on realistic engineering tasks.

Software Development Intern - Bluestock Fintech - Remote

(2025-05 - 2025-07)

  • Developed full-stack IPO platform with Django, PostgreSQL, and Django REST Framework — secure REST APIs, RBAC admin dashboard, JWT authentication, and advanced search/filtering over IPO metadata.

Education

B.Tech - Mechanical Engineering - Indian Institute of Technology (ISM), Dhanbad (2022 - 2026)

Skills

AI agent testingand OCR/document-AI QA. Open to remote roles.

Reviews

Related