Vasyl T

Senior Software Engineer

myops • Temps plein

Oct 2026 - Present • 1 mo

I am working on AI Powered Real Estate CRM.

Lead Software AI Engineer

Intrepide • Temps plein

Jan 2026 - Oct 2026 • 9 mos

▸ Lead a small engineering team building the backend of an AI agent platform (Python, FastAPI): own architecture decisions, technical roadmap, code review, and delivery. ▸ Design and ship production-grade agentic systems with LangGraph on top of Amazon Bedrock models: multi-agent orchestration, long-running multi-step workflows, tool use, and retrieval over enterprise knowledge bases. ▸ Build MCP connectors and servers that expose internal and third-party systems as tools for agents, and author MCP metadata schemas aligned with Canonical Contract definitions for consistent tool discovery. ▸ Designed the platform's agent and tool registry on top of the AWS agent registry: data models covering contracts, metadata, versioning, and lifecycle states (draft → active → deprecated → retired), with agent registration and discovery exposed via REST APIs with strict Pydantic validation. ▸ Develop data-intensive backend services and RAG pipelines (PostgreSQL / pgvector, Redis, vector stores) that power agent retrieval and context management. ▸ Own quality and safety of LLM systems: agent and LLM evaluation pipelines, regression testing across prompt and model changes, and LLM security (prompt-injection defense, guardrails, scoped tool access). ▸ Implement authentication and authorization (JWT, OAuth2, role-based access) for internal and external APIs. ▸ Set engineering standards for the team: API design conventions, testing, observability, and review process

Middle Python Developer

DTEAM • Temps plein

Feb 2025 - Dec 2025 • 10 mos

Developed data-intensive backend services and REST APIs for client projects, handling large datasets with PostgreSQL and Redis-backed caching and background processing. ▸ Built AI agents and LLM-powered features with LangChain/LangGraph, including RAG systems with retrieval pipelines, vector stores, and knowledge bases. ▸ Designed and ran LLM regression test suites to catch quality drift across prompt, model, and pipeline changes. ▸ Performed LLM security work: prompt-injection and jailbreak testing, input/output guardrails, and hardening of agent tool access. ▸ Wrote automated tests and API documentation; participated in code reviews and sprint planning