Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

I’m a software engineer with nearly four years of experience spanning AI/ML applications, full-stack development, cloud infrastructure, and distributed systems. I’ve built AI-enabled workflows, scalable backend services, responsive web applications, and cloud-based platforms aimed at production-grade reliability and security. My focus areas include practical ML evaluation, RAG-based systems, API design, observability, system reliability, and performance optimization to deliver robust, enterprise-ready AI software solutions.…I’m a software engineer with nearly four years of experience spanning AI/ML applications, full-stack development, cloud infrastructure, and distributed systems. I’ve built AI-enabled workflows, scalable backend services, responsive web applications, and cloud-based platforms aimed at production-grade reliability and security. My focus areas include practical ML evaluation, RAG-based systems, API design, observability, system reliability, and performance optimization to deliver robust, enterprise-ready AI software solutions.

Pavan Vishnu Sai Bestha

Full Stack Developer, AI Developer, Web Developer, +2





I’m a software engineer with nearly four years of experience spanning AI/ML applications, full-stack development, cloud infrastructure, and distributed systems. I’ve built AI-enabled workflows, scalable backend services, responsive web applications, and cloud-based platforms aimed at production-grade reliability and security. My focus areas include practical ML evaluation, RAG-based systems, API design, observability, system reliability, and performance optimization to deliver robust, enterprise-ready AI software solutions.…I’m a software engineer with nearly four years of experience spanning AI/ML applications, full-stack development, cloud infrastructure, and distributed systems. I’ve built AI-enabled workflows, scalable backend services, responsive web applications, and cloud-based platforms aimed at production-grade reliability and security. My focus areas include practical ML evaluation, RAG-based systems, API design, observability, system reliability, and performance optimization to deliver robust, enterprise-ready AI software solutions.

Available to hire

I’m a software engineer with nearly four years of experience spanning AI/ML applications, full-stack development, cloud infrastructure, and distributed systems. I’ve built AI-enabled workflows, scalable backend services, responsive web applications, and cloud-based platforms aimed at production-grade reliability and security.

My focus areas include practical ML evaluation, RAG-based systems, API design, observability, system reliability, and performance optimization to deliver robust, enterprise-ready AI software solutions.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Language

English

Fluent

Work Experience

AI Software Engineer at Mercor – Agentic AI Evaluation & Reliability

November 1, 2025 - Present

Engineered internal tooling for LLM-based coding agents, supporting multi-step agent workflows, execution pipelines, validation checks, regression artifacts, and reviewer-facing debugging summaries across 100+ evaluation runs. Implemented multi-signal quality gates evaluating outputs across correctness, instruction following, runtime behavior, reproducibility, hallucination risk, code quality, and safety across 6+ dimensions. Defined failure taxonomies for agentic workflows across 10+ categories. Strengthened observability by analyzing logs, traces, build outputs, test failures, runtime errors, tool calls, and changed-file summaries. Evaluated AI agent workflows across large Next.js, Python, and Node.js codebases, improving reliability and pass rates to ~95%. Captured reviewer judgments and quality signals to support human-in-the-loop feedback, evaluation datasets, and fine-tuning workflows. Refined prompts, retrieval context, tool usage, configurations, and execution logic to reduce r

Senior AI/ML Evaluation Engineer at Coursera

December 1, 2025 - March 1, 2026

Validated AI/ML platform workflows across 10+ course modules covering microservice-based inference, reproducible experimentation, and production-grade ML evaluation practices. Built AI-graded labs simulating Kafka-style ingestion, batched workers, throughput/lag evaluation, and asynchronous ML system behavior for learners. Authored grading and validation rubrics across 4 evaluation areas, including ablation studies, reproducibility, seed control, variance interpretation, and fair model comparison.

Full-Stack Developer at Cinema Verde

August 1, 2024 - August 1, 2025

Re-architected a media and analytics platform using React, Node.js, Java Spring Boot, PostgreSQL, and AWS, improving page load time by 35%, visitor retention by 30%, and platform uptime to 99.9%. Migrated large media assets from MySQL BLOBs to Amazon S3 with metadata in PostgreSQL, boosting scalability and query performance. Engineered an event-driven AWS media pipeline (S3-Lambda-MediaConvert) to automate video processing for 50+ videos. Integrated an AI chatbot for content discovery using RAG over film records, enhancing search and exploration; improved chatbot quality via prompt tuning and retrieval filtering. Mentored 5 engineers, increasing sprint velocity by 25% and reducing handoff time by 20%.

Software Engineer at I Ray IT Solutions

January 1, 2022 - December 1, 2022

Architected C# ASP.NET Core REST APIs handling 50k+ daily sensor events with peak loads of 2k/min and sub-200ms latency. Built backend services for real-time traffic monitoring, improving data ingestion and reliability. Containerized microservices with Docker and provisioned automated AWS EC2 load-testing environments. Implemented CloudWatch logs, metrics, and SES alerts, reducing issue detection time from hours to under 5 minutes.

Frontend UI Developer at Department of Mechanical MUJ

August 1, 2021 - December 1, 2021

Prototyped an 8-semester planning web tool with WCAG-aligned accessibility. Conducted user research with 50+ users, translating feedback into wireframes and reusable UI patterns. Delivered a responsive high-fidelity prototype, accelerating stakeholder approvals and improving developer handoff quality.

ML Researcher at BITS Pilani Hyderabad Campus

January 1, 2021 - July 1, 2021

Delivered an Arduino-based height estimation system with optimized signal capture and preprocessing. Curated datasets, trained ML models, and validated results using repeatable evaluation workflows to improve prediction reliability. Optimized OpenCV preprocessing for real-time computer vision experiments under latency constraints.