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

I’m an AI-native full-stack engineer who started building as soon as AI-assisted development became practical. After three years, I have 4 systems in production for real businesses and 25+ projects spanning POS/billing, AI products, and ML infrastructure. I ship end-to-end—from schema to deployment—then iterate against real counter use. I focus on evaluation, reliability, and production-grade engineering, including adversarial benchmark harnesses, LLM/RAG/finetuning pipelines, and secure multi-tenant systems.…I’m an AI-native full-stack engineer who started building as soon as AI-assisted development became practical. After three years, I have 4 systems in production for real businesses and 25+ projects spanning POS/billing, AI products, and ML infrastructure. I ship end-to-end—from schema to deployment—then iterate against real counter use. I focus on evaluation, reliability, and production-grade engineering, including adversarial benchmark harnesses, LLM/RAG/finetuning pipelines, and secure multi-tenant systems.

Boopathi Raja

AI Engineer, AI Developer, Full Stack Developer, +4





I’m an AI-native full-stack engineer who started building as soon as AI-assisted development became practical. After three years, I have 4 systems in production for real businesses and 25+ projects spanning POS/billing, AI products, and ML infrastructure. I ship end-to-end—from schema to deployment—then iterate against real counter use. I focus on evaluation, reliability, and production-grade engineering, including adversarial benchmark harnesses, LLM/RAG/finetuning pipelines, and secure multi-tenant systems.…I’m an AI-native full-stack engineer who started building as soon as AI-assisted development became practical. After three years, I have 4 systems in production for real businesses and 25+ projects spanning POS/billing, AI products, and ML infrastructure. I ship end-to-end—from schema to deployment—then iterate against real counter use. I focus on evaluation, reliability, and production-grade engineering, including adversarial benchmark harnesses, LLM/RAG/finetuning pipelines, and secure multi-tenant systems.

Available to hire

I’m an AI-native full-stack engineer who started building as soon as AI-assisted development became practical. After three years, I have 4 systems in production for real businesses and 25+ projects spanning POS/billing, AI products, and ML infrastructure.

I ship end-to-end—from schema to deployment—then iterate against real counter use. I focus on evaluation, reliability, and production-grade engineering, including adversarial benchmark harnesses, LLM/RAG/finetuning pipelines, and secure multi-tenant systems.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Intermediate

Beginner

Language

Tamil

Advanced

English

Advanced

Work Experience

AI Evaluation Engineer at Project Dynamo Handshake AI (Contract)

July 1, 2026 - Present

Built adversarial terminal benchmark tasks and evaluation pipelines for frontier-agent capability measurement. Designed task specs, references/automations, and independent simulation verifiers in Python/Docker/pytest; validated difficulty with trial runs and wrong-solver resilience. Automated review/retrieval pipeline and produced evidence-based evaluation artifacts used for real-world testing, deployment readiness, and ongoing calibration.

AI Evaluation Engineer — Coding-Agent Benchmarks at Handshake AI

January 1, 2026 - Present

Authored adversarial command-line benchmark tasks for Terminal-Bench 2 / Harbor to measure frontier coding-agent capability. Built maintainer-style task specs, reference solutions, and sealed verifier suites in Python/Docker/pytest. Engineered hidden tests to be non-gameable using wrong-solver batteries, spec-prose-based blind solver agreement checks, and three-way fuzzing to find divergences. Validated difficulty with pass@5 trials against a published gate and ensured tasks cleared automated review stages at a 0/5 agent solve rate. Submitted on a single push with full green automated checks.

Independent Full-Stack & AI Engineer (Freelance)

January 1, 2023 - Present

Built and maintained multiple client systems in daily production use, shipping end-to-end features and post-launch iterations. Worked on Steel Flow (steel-trading billing & inventory), Women’s Zone (variant-aware retail POS), The Signature (bakery POS with kitchen-order printing), and Prepli (exam-prep platform for Indian government-exam aspirants). Developed Nexq (Australian digital queue & booking platform) with Stripe subscription billing, resumable OTP onboarding, multi-tenant row-level security across migrations, edge functions, Playwright end-to-end specs, and CI auto-deploy. Delivered Prism (Power BI request-intake portal) with role-gated Postgres RLS, Azure AD client-credentials integration, and automated email/Teams notifications with nightly E2E runs against production.

Independent Full-Stack & AI Engineer at Freelance

January 1, 2023 - Present

Developed and shipped multiple end-to-end AI systems and production/MVP/prototype projects across POS/billing, AI products, and ML infrastructure. Implemented LLM integrations (Claude/OpenAI/Gemini), agents/tool calling, RAG, QLoRA fine-tuning pipelines, ONNX/vLLM/edge inference workflows, CI/CD and infrastructure automation. Worked on real deployments including queue/booking, role-gated workflows, BI portals, fraud/evidence platforms, and on-device quantized LLM runtimes with test suites.