Looking for a job

Petr Molchanov

@petrmolchanov

Senior AI Engineer at ROGII building a production multi-agent coding system that reviews fixes and opens merge requests.

Argentina

Message

At ROGII, I built and operate an autonomous multi-agent coding system that picks up Jira tasks, writes fixes, passes independent AI review, and opens merge requests. I also built its guardrails, RAG layer over project documentation and code history, and evals for measuring quality.

I moved QA from manual work to agent-driven Playwright end-to-end testing, and created reusable agent skills that support onboarding across QA, product, and engineering.

As an independent engineer, I built LLM pipelines and more than 20 AI features for a B2B parts marketplace, using structured outputs, tracing, and an eval suite over 1,300+ labeled examples. I also built a read-only MCP server and AI analytics for a labor-planning SaaS.

Before focusing on AI, I worked on frontend and full-stack products. At ROGII, I built a Canvas-based well cross-section visualization and migrated the mapping layer to Mapbox; at ROKO Labs, I reduced bundle size by 15% and page load time by 20%, and built a virtualized grid for 10,000+ rows at 60fps.

Experience

Work history, roles, and key accomplishments

Current

Senior AI Engineer

Current

ROGII

Nov 2025 - Present (1 year)

• Built and operate an autonomous multi-agent coding system in production (Claude Agent SDK): it picks up Jira tasks, writes the fix, passes independent AI review and opens merge requests across frontend, backend and e2e test repositories.
• Added human-in-the-loop guardrails: a risk classifier decides which changes need human approval.
• Built Retrieval-Augmented Generation (RAG) over project doc

Independent AI & Full-Stack Engineer

Self-Employed

Sep 2024 - Aug 2026 (1 year 11 months)

Part-time side projects alongside full-time work: LLM pipelines, AI features and MCP servers.
• Built a tiered LLM pipeline for a B2B parts marketplace (NestJS, PostgreSQL/pgvector, Python ML service): rules first, then local classifiers (multilingual-e5), then a hosted LLM with OpenRouter fallback.
• Shipped 20+ AI features with Zod-validated structured outputs, per-feature USD budgets, Langfuse

Software Engineer

ROKO Labs

Jan 2022 - Jan 2026 (4 years)

Long-term contract engagement (renewed annually).

• Refactored a large React/TypeScript app, cutting bundle size 15% and page load time 20%.
• Built a virtualized data grid for 10,000+ rows at 60fps.
• Implemented Redux-Observable middleware for complex async data flows, reducing boilerplate by ~30%.
• Worked with backend engineers on versioned REST API design, keeping backward compatibility acro

Senior Software Engineer

ROGII

Jun 2023 - Oct 2025 (2 years 4 months)

Core UI of a cloud platform for managing and visualizing oil & gas well data across global multi-project operations (React, TypeScript). Teams use it to monitor and analyze wells across entire portfolios, not single sites.

• Built a Canvas-based 2D cross-sectional visualization for well spacing and collision analysis, shipped as part of the platform.
• Migrated the production mapping layer from L

Full Stack Software Engineer

LLC PROFSOFT

Jun 2020 - Jan 2022 (1 year 7 months)

• Built a React/Redux e-learning platform serving 500+ users: course management, onboarding flows and AWS S3 video uploads.
• Designed and implemented NestJS backend modules for document workflows and file management.
• Raised unit and integration test coverage from 45% to 78% with Jest.
• Configured GitHub Actions CI/CD pipelines for automated testing, builds and AWS deployments.