Agentic Engineer
About Octocrew
Octocrew builds and runs managed AI coworker crews — small teams of specialist AI agents that work inside a company's existing tools and help human teams execute recurring operational and marketing workflows. We are expanding our contractor pool with practical engineers who can implement, integrate, tune, and support client AI agent workflows.
The role in one line
You take an approved requirements spec and turn it into a working agent system that fits the client's tools, data, permissions, and approval process. This is not a prompt-only role and not pure DevOps; it sits between backend engineering, automation, LLM application development, systems integration, and client implementation. You do not need to already know our full agent stack — we have templates and patterns, and we care more about whether you can become reliable fast.
What you will do
Translate a client workflow into a clear implementation plan: tools, data sources, permissions, approval points, edge cases, success criteria.
Build Python glue around AI workflows: API connectors, webhooks, small backend services, scripts, data syncs, and browser automation where APIs are limited.
Configure and tune LLM-powered workflows from our templates.
Integrate Slack, Telegram, Notion, GitHub, CRMs, ecommerce and marketing platforms, helpdesks, analytics, and internal knowledge bases.
Make implementations maintainable: logs, error handling, smoke tests, runbooks, handoff notes.
Support launched workflows: triage issues, debug integrations, fix regressions, and document changes.
What we expect
2+ years of backend, automation, integration, or applied software engineering experience.
Python-first engineering ability.
Solid API fundamentals: REST, webhooks, OAuth basics, retries, rate limits, request debugging, structured logs.
Some hands-on exposure to LLM tools or AI automation, or strong motivation to learn them quickly with guidance.
Comfort working with messy real-world systems: unclear requirements, imperfect docs, partial APIs, edge cases.
Basic DevOps literacy: Linux, Docker basics, environment variables, logs.
Pragmatic scoping and a self-directed work style.
Ukrainian for team communication; working English for documentation and APIs.
Nice to have
Integrations experience (Slack, Telegram, Notion, GitHub, Google Workspace, CRM, helpdesk, ecommerce, analytics); OpenAI / Anthropic / Gemini, RAG, vector search, or agent orchestration; Playwright; client-facing experience; security and data-handling awareness.
Format
Contractor / project-based, remote. Start with a compact paid pilot task that mirrors real client work, then project delivery and/or ongoing support. Compensation hourly or project-based, agreed after technical fit and scope.