AI Software Engineer

We are looking for a results-driven Backend Engineer to design and build automation infrastructure for our clients business workflows. In this role, you will bridge the gap between deterministic backend logic and probabilistic AI models. You will be responsible for orchestrating complex data flows, integrating with third-party APIs, and ensuring that our LLM-driven features are reliable, scalable, and fault-tolerant in a production environment.

Core Responsibilities

• Workflow Orchestration: Design and maintain the backend architecture for multi-step automation chains that ingest user intent, query internal databases, and trigger actions in external systems.
• LLM & API Integration: Build secure, efficient connectors for LLM providers (OpenAI, Anthropic) and third-party SaaS platforms. Implement function calling and tool-use patterns to turn natural language into structured API requests.
• System Resilience: Engineer “defensive” logic to handle the non-deterministic nature of AI. Implement robust retry strategies, rate-limiting handling, and fallback mechanisms to ensure system stability despite model hallucinations or latency.
• Prompt Engineering & Ops: Write, test, and version-control system prompts. Optimize context windows and enforce structured outputs (e.g., JSON schemas) to guarantee downstream compatibility.

The Stack & Requirements

• Backend Proficiency: 4+ years of production experience in Python (FastAPI/Django), TypeScript (Node.js), or other strongly typed languages. Strong command of asynchronous programming and RESTful API design.
• LLM Experience: Practical experience building applications with LLMs. You understand the nuances of temperature, token limits, and how to prevent prompt injection.
• Data & Storage: Proficiency with SQL/NoSQL databases and experience with Vector Stores (Pinecone, Weaviate) for context retrieval.
• Error Handling: Strong habit of using validation libraries (e.g., Pydantic, Zod) to sanitize and verify AI-generated data before it touches business logic.

Nice to Haves

• Experience with RAG (Retrieval-Augmented Generation) pipelines.
• Familiarity with distributed task queues (Celery, BullMQ, Kafka).
• Background in building integrations for platforms like Slack, Salesforce, or GitHub.