CTO, co-founder, architect, and hands-on full-stack engineer.

I build AI-native systems end-to-end: product architecture, backend, frontend, infra, testing, and rollout.

• agentic systems and execution runtimes
• agentic RAG, semantic search, and organization-aware context
• AI infrastructure, evals, and testing
• governance layers for humans and AI agents
• structured AI interfaces with AG-UI / A2UI patterns

• I work not only with models and prompts, but with operating systems for teams
• I translate self-management patterns from Sociocracy 3.0 / Holacracy-inspired governance into executable agent behavior
• I care about permissions, approvals, audit, handoffs, and clear responsibility boundaries
• I build separate execution layers for agents, including Codex SDK-based runner patterns, instead of collapsing orchestration into one app process
• I treat retrieval as a product surface: hybrid search, agentic RAG loops, and ZVec-backed semantic search with clear operational boundaries

• currently CTO and co-founder of Anavi
• still hands-on: I write core product code and lead engineering / architecture
• the product is shifting from active build-out toward support mode for existing customers, so I am open to the next role

• organization-scoped agentic RAG with hybrid retrieval and semantic search
• separate agent runner / executor layer with Codex SDK-backed execution in the main product
• semantic search accelerated with ZVec, with pgvector-compatible fallback paths
• rich AI result rendering with AG-UI / A2UI patterns for structured answers

• CTO
• Founding Engineer
• Staff / Principal Engineer
• Senior Full-stack / Backend roles in AI-native companies

Formats:

• full-time
• contract
• part-time
• remote / hybrid / office
• relocation

• LinkedIn: https://www.linkedin.com/in/agritsev/
• Telegram: https://t.me/agrisevich
• Website: https://gritsevich.com
• Main product: https://anavi.team

These public repositories are public-safe extractions from production-backed work around agent governance, execution, retrieval, and rich answers, not side experiments.

• org-aware-agents: architecture notes on roles, policies, consent, coordination, and org-aware agent surfaces
• agent-code-playbook-kit: process IR and deterministic compilation into runtime playbooks
• agent-code-runner-sandbox: minimal runner patterns for sandboxing, locks, and delivery validation
• agent-code-observer-mcp: semantic MCP surface and thin observer bridge patterns
• agent-code-a2ui-contracts: direct A2UI v0.8 contracts, normalization, and schema helpers for rich agent answers
• gritsevich-com: personal site and bilingual writing shell

Together these repositories cover:

• governance-first agent architecture
• process authoring and deterministic runtime playbooks
• bounded execution and runner patterns
• shared semantic tool surfaces for internal and external agent clients
• structured final-answer contracts for AG-UI + A2UI style interfaces
• public writing and product framing

If you are building AI-native systems and need someone who can combine product sense, architecture, governance, and hands-on implementation, feel free to reach out.