Open to opportunities

Rishav Kumar

@rishavkumar6

Message

I build reliable AI and backend systems, focused on evaluation and benchmarks.

India

Message

What I'm looking for

I’m looking for a role where I can build dependable AI-backed products and backend services, and also strengthen evaluation/benchmarking—shipping improvements that raise reliability, accuracy, and safety in real-world workflows.

I’m a computer science and engineering student building AI systems and backend services with a strong emphasis on evaluation quality. I’ve designed real-world benchmark tasks to assess frontier AI coding agents on debugging, API integration, and code maintenance workflows, using Docker-based environments for reproducibility.

In engineering roles, I optimized LLM prompt pipelines for Python code generation and improved reliability by 35% while cutting manual debugging time by 40%. I also delivered backend features like RESTful APIs processing 1,000+ call workflows and implemented a regex-based PII masking system that redacts 95%+ sensitive data, alongside RAG chatbot work that improved query response accuracy by 45% and reduced deployment time by 30%.

Experience

Work history, roles, and key accomplishments

Current

AI Coding Evaluation Contributor

Current

Snorkel AI

Jan 2026 - Present (6 months)

Designed and developed real-world software engineering benchmark tasks to evaluate frontier AI coding agents across debugging, API integration, system configuration, and code maintenance. Created oracle solutions and automated Docker-based test suites, then iteratively refined benchmarks using evaluation and peer-review feedback.

Benchmarking Debugging API Integration System Configuration Software Maintenance Test Suites Multi Step Reasoning Code Agent Assessment

AI Engineer

AirDwagLabs

Apr 2026 - Jun 2026 (2 months)

Optimized LLM prompt pipelines for Python code generation using multi-turn debugging and iterative refinement to improve code reliability and reduce manual debugging time. Conducted end-to-end evaluation of AI coding agents and contributed to model assessment workflows by defining benchmarks and identifying misleading outputs.

Multi Turn Debugging AI Coding Agent Evaluation Response Quality Analysis Benchmarking

SDE Intern

Xelron.AI

Jan 2026 - Mar 2026 (2 months)

Developed Python-based AI systems by identifying key failure cases to improve AI agent reliability and consistency. Refined RLHF workflows through review of 100+ model outputs and provided actionable feedback to improve response accuracy and overall AI evaluation processes.

Python AI Systems response evaluation Feedback Workflows

SDE Backend Developer Intern

Futwork

Aug 2025 - Dec 2025 (4 months)

Developed three RESTful APIs (Call Script, Transcript, and Analysis) using Gemini AI to process 1,000+ AI-user call workflows and improve call-stage detection accuracy. Built a regex-based PII masking system and enhanced backend automation for AI-powered call analysis with secure preprocessing to reduce data privacy risks.

Python REST APIs Gemini AI Script Based Calling Workflow Analysis PII Masking Data Privacy

Summer Research Intern

Indian Space Academy

Jun 2025 - Jul 2025 (1 month)

Conducted data-driven astronomy research using real-world datasets to estimate galaxy cluster masses. Used Python and machine learning with data preprocessing, visualization, and statistical analysis to improve interpretation accuracy.

Python Machine Learning data preprocessing Visualization Statistical Analysis Data Driven Astronomy Galaxy Cluster Mass Estimation