Tech Lead for AI Agents (Production-Grade Development)

Alongside partners with organizations to drive digital transformation and build high-performing tech teams.

We’re looking for a hands-on Tech Lead to build production-grade AI agents and lead delivery from architecture to deployment. You’ll code, make technical decisions, and help the team ship reliable AI-powered products quickly.

Responsibilities:

• Build agent systems with Hermes Agent, OpenClaw, Letta, Agent Zero, OpenHands, or similar open-source runtimes.
• Use frameworks such as OpenAI Agents SDK, LangGraph, Google ADK, Vercel AI SDK, CrewAI, AutoGen, or Mastra.
• Develop reusable skills, tools, subagents, workflows, persistent memory, scheduled tasks, and messaging integrations.
• Integrate OpenAI, Anthropic, Gemini, open-weight, and local models through tool calling, structured outputs, streaming, and model routing.
• Connect agents to systems using MCP, A2A, REST APIs, webhooks, browsers, terminals, and computer-use tools.
• Build RAG and context systems using pgvector, Pinecone, Weaviate, Elasticsearch, or equivalent technologies.
• Establish evaluations and observability with LangSmith, Braintrust, Phoenix, OpenTelemetry, or provider tracing.
• Secure autonomous agents using sandboxing, least-privilege access, approval gates, secrets management, audit trails, and prompt-injection defenses.
• Use Codex, Claude Code, Cursor, or GitHub Copilot to accelerate delivery.
• Own technical plans, production quality, latency, reliability, and model costs.

Requirements

• Strong software engineering and hands-on technical leadership experience.
• Experience shipping AI agents or LLM applications into production.
• Strong Python or TypeScript skills and cloud infrastructure experience.
• Practical knowledge of agent orchestration, memory, tool use, RAG, evaluations, and human-in-the-loop design.
• Familiarity with current trends such as self-improving agents, reusable agent skills, local-first and bring-your-own-model runtimes, multimodal and computer-use agents, MCP/A2A interoperability, durable execution, and eval-driven development.