Software Architect – AI Security Platform

About the role

You will design and own a sophisticated, multi‑tenant AI security platform that delivers a 7‑stage LLM‑powered control‑generation pipeline. The solution is production‑grade, requiring reproducibility, security, and auditability, and you will guide a small team of AI engineers through high‑level design decisions.

Key responsibilities

• Act as the architecture authority – define the deterministic/AI boundary, multi‑tenant isolation model, and LLM security posture.
• Make baseline architectural decisions in the first week that lock the team’s constraints.
• Own the end‑to‑end delivery using FastAPI, PostgreSQL, AWS Bedrock and ECS Fargate while preventing scope creep.
• Partner with client leadership to translate business and compliance requirements into architectural constraints.
• Evaluate technology trade‑offs such as single‑step AI processing, row‑level security versus application‑layer filtering, and Pydantic validation.
• Identify and mitigate delivery risks, including prompt‑injection surfaces and cross‑tenant isolation defects.
• Define long‑term product roadmap, technical vision, and development strategy.
• Review critical code paths, create architectural decision records, and ensure documentation captures the “why”.
• Lead, mentor, and support the AI engineering team.

Required profile

• Proven ability to lead and mentor a small team of AI engineers.
• Strong decision‑making skills for high‑level architectural challenges.
• Excellent communication with stakeholders to translate business and compliance needs into technical solutions.

Required skills

• Python
• FastAPI
• PostgreSQL
• AWS (Bedrock, ECS Fargate)
• LLM integration and prompt‑injection mitigation
• Multi‑tenant data isolation (Row Level Security)
• Pydantic schema validation