Senior Software Engineer - AI

Role overview

What you will do

• Select practical approaches across LLMs, RAG, tool calling, agents, deterministic workflows, and traditional software.
• Build production AI services with Python, FastAPI, asynchronous workers, PostgreSQL, Redis, queues, model APIs, and external tools.
• Build tool-calling agents and multi-step workflows that interact with internal APIs, MCP tools, business systems, and knowledge sources.
• Design event-driven systems with retries, dead-letter handling, idempotency, back pressure, and failure recovery.
• Build and improve RAG systems including ingestion, chunking, embeddings, retrieval, reranking, metadata filtering, context construction, and citations.
• Evaluate models and diagnose hallucinations, retrieval failures, agent issues, latency, provider failures, and inference costs.
• Create AI evaluation pipelines using datasets, regression tests, retrieval metrics, groundedness checks, and tool execution evaluation.
• Own implementation across APIs, workflows, data, queues, model integration, evaluations, observability, and production support.

What we are looking for

• 5 or more years of professional software engineering experience building production backend, distributed, or cloud-based systems.
• Production experience building Applied AI, LLM, RAG, NLP, or agent-based applications.
• Advanced Python, FastAPI, asynchronous programming, Pydantic, SQLAlchemy or SQLModel, and production API development skills.
• Strong backend and distributed systems fundamentals including REST APIs, concurrency, caching, reliability, and production debugging.
• Production experience with RabbitMQ, Kafka, Azure Service Bus, or equivalent messaging architectures.
• Experience integrating LLM APIs and building structured output, function calling, tool calling, or agent execution workflows.
• Experience building a RAG or knowledge-grounded system with measurable quality and latency outcomes.
• Strong PostgreSQL and data modeling skills, plus Redis, pgvector, vector databases, or hybrid search experience.
• Demonstrated daily use of AI coding agents such as Cursor, Claude Code, Codex, or equivalent tools.

Nice to have

• Understanding of prompting, structured outputs, tool schemas, context management, model routing, retries, fallbacks, rate limits, and streaming responses.
• Ability to design measurable experiments and evaluations.
• Understanding of retrieval quality, grounding, hallucination mitigation, prompt injection risks, tool authorization, and agent guardrails.
• Experience with Docker, CI/CD, automated testing, observability, distributed tracing, OpenTelemetry, Langfuse, or similar tools.
• Experience with Semantic Kernel, Microsoft Agent Framework, LangGraph, PydanticAI, or equivalent agent frameworks.
• Experience with Model Context Protocol tools or servers, LiteLLM, Azure AI Search, pgvector, Chroma, Kubernetes, Azure, OCI, or similar technologies.
• Experience with high-scale customer-facing AI products, open-weight model inference, AI safety controls, or automated agent evaluation systems.

How you will work

• Independently break down ambiguous product goals, validate assumptions, choose an approach, and drive solutions through production.
• Choose the simplest architecture that solves the problem rather than using an agent or LLM by default.
• Use AI coding agents to improve engineering velocity while preserving quality, testing, security, and maintainability.