AI Workflow & Autonomous Agent Architect

At Cehpoint, we build commercial AI automation systems that perform real work. As an AI Agent Architect, you will design, test, and deploy resilient autonomous agents capable of web browsing, document synthesis, database querying, and cross-system API execution.

• Design and deploy multi-agent collaboration frameworks using state machines, feedback loops, and fallback guards.
• Implement robust function calling and structured JSON tool-use pipelines across Gemini, Claude, and open-source models.
• Build production RAG pipelines integrating dense semantic search, hybrid re-ranking, and pgvector knowledge bases.
• Optimize model latency, token budgets, caching layers, and multi-tenant rate limiting.
• Continuously evaluate agent behavior against hallucination benchmarks, safety policies, and regression test suites.

• Deep understanding of LLM capabilities, attention mechanisms, context limits, and tokenization trade-offs.
• Familiarity with agentic design patterns: ReAct, Plan-and-Solve, Self-Refine, and Multi-Agent consensus.
• Strong computer science fundamentals: async concurrency, queuing architectures, and distributed systems.

• Extensive hands-on Python (FastAPI/AsyncIO) and/or Node.js modern backend development.
• Experience integrating OpenAI, Anthropic, Gemini, or OpenRouter APIs with complex tool-calling schemas.
• Expertise with vector indexing (pgvector, Qdrant, Chroma) and embedding generation.
• Proficiency in containerized deployment using Docker and Linux server administration.

• Obsession with cutting-edge open-source LLM research, agentic frameworks, and autonomous tooling.
• A builder mindset focused on creating autonomous workflows that solve real commercial problems.
• Commitment to writing self-healing, deterministic agent code with rigorous error boundaries.

• Premium compensation (₹10,00,000–₹18,00,000) with annual review and equity participation options.
• High-velocity startup culture with zero red tape and immense technical autonomy.
• Full access to state-of-the-art GPU clusters, frontier LLM API credits, and research tooling.
• Work from anywhere in India with flexible scheduling.