AI Platform Engineer – Autonomous Agent Systems

About the role

PSTAG is seeking an AI Platform Engineer to design and deploy advanced autonomous agent systems. You will shape AI‑driven solutions that integrate with internal and external services, working remotely from Costa Rica.

Key responsibilities

• Design and implement autonomous agent architectures using frameworks such as Hermes Agent.
• Develop multi‑agent collaboration patterns (orchestrator‑workers, debate, hierarchical swarms).
• Build agentic memory systems with short‑term, long‑term, and episodic storage using vector databases and semantic caching.
• Integrate advanced reasoning techniques including ReAct, Chain‑of‑Thought, Tree‑of‑Thoughts, and Plan‑and‑Solve.
• Enable dynamic planning, error recovery, and replanning based on environmental feedback.
• Implement tool use and API grounding for database queries, RAG retrieval, UI automation, and other actions.
• Create robust evaluation frameworks to assess task completion, efficiency, and safety.
• Instrument agents with tracing, observability, and logging tools such as LangSmith, Arize, and Weights & Biases.
• Optimize agents for latency, token cost, and production reliability.
• Integrate agents with CRMs, databases, Slack, browsers, REST APIs, and code interpreters.

Required profile

• Minimum 3 years of professional software engineering experience.
• Strong expertise in Python programming.
• Conversational proficiency in English.
• Demonstrated experience building or prototyping agentic patterns (ReAct, Reflexion, Toolformer) in production.

Required skills

• Python
• Hermes Agent framework
• Vector databases (Pinecone, Weaviate, Qdrant)
• LangSmith, Arize, Weights & Biases
• Prompt engineering and few‑shot learning
• Structured output generation (JSON mode, grammars)
• ReAct, Reflexion, Toolformer
• Chain‑of‑Thought, Tree‑of‑Thoughts, Plan‑and‑Solve
• Integration with CRMs, Slack, REST APIs, and code interpreters