Senior Full Stack Engineer (AI)

Responsibilities:

• AI Platform and Pipelines: Design and build production RAG systems, agentic workflows, and LLM orchestration layers. Pick the right models, chain them, ground them.
• Evals and Model Quality: Build rigorous evaluation frameworks. Define metrics that matter, run systematic benchmarks, and continuously improve model accuracy, latency, and cost.
• Scalable Infra: Architect backend services (APIs, queues, vector stores) that handle millions of legal documents without falling over. Own deployment, monitoring, and cost optimisation on cloud (AWS/GCP).
• Product Engineering: Ship user-facing features with React/Next.js . You care about what the attorney actually sees and uses, not just what the model outputs.
• Data and Embeddings: Work with massive structured and unstructured legal datasets. Build and maintain embedding pipelines, chunking strategies, and retrieval infrastructure.
• MLOps and Lifecycle: Own the full model lifecycle: fine-tuning, deployment, A/B testing, monitoring, and rollback.

Requirements:

• 5+ years shipping production software; strong Python plus at least one of Go, Rust, or Kotlin.
• Hands-on LLM experience: you've built real applications against OpenAI, Anthropic, or open- source models (not just called an API once).
• Deep understanding of RAG architectures, prompt engineering, embeddings, and retrieval systems.
• Solid backend engineering: APIs, microservices, distributed systems, SQL/NoSQL databases.
• Cloud deployment experience (AWS/GCP/Azure), you've scaled services, not just deployed demos.
• Strong CS fundamentals: data structures, algorithms, system design.
• Frontend proficiency: React.js / Next.js or equivalent. You can build an entire AI application, not just a backend.

Bonus points:

• Experience with vector databases (Pinecone, Weaviate, FAISS, Qdrant).
• MLOps tooling: MLflow, Airflow, Kubeflow, or similar.
• Real-time data pipelines or streaming systems.