Senior AI Engineer

AI Builder Track · Hands-on IC · High Bar

The Role

You build the AI systems behind our safeguarding platform: LLM-powered pipelines that read large volumes of social media content and surface real risk, fast, accurately, and explainably. This is production AI engineering, not research. You ship, you measure, you own it.

What You'll Do

• Design and ship LLM-powered services for classification, extraction, and risk detection over real-time and batch data
• Build agentic workflows: multi-step reasoning, tool use, structured outputs, and guardrails
• Own evals: build eval sets, quality metrics, and LLM-judge pipelines that catch regressions before customers do
• Build retrieval pipelines (embeddings, hybrid search, reranking) and tune them for precision and recall
• Make trade-offs between accuracy, latency, and cost per check, and document why
• Instrument AI services end to end: tracing, prompt and version tracking, drift and failure monitoring
• Own your systems in production, including on-call and incident response

What You Bring

• 5+ years building production software, including 2+ years shipping LLM-based systems to real users
• Strong Python and/or TypeScript, plus backend fundamentals (APIs, queues, databases, caching)
• Hands-on experience with frontier model APIs (Claude, OpenAI, xAI) and agent/LLM frameworks
• A track record of designing evals and improving quality with numbers, not vibes
• Understanding of LLM failure modes (hallucination, prompt injection, bias) and how to guard against them
• AI-native development: you write most of your code using AI, but you own every line that ships

Nice to Have

• RAG at scale and hands-on experience with vector databases
• Multimodal (image/video) models
• Trust & safety or content moderation experience

Who This Role Is NOT For

• Engineers who treat a demo as done: we measure in production
• Those who see on-call as someone else's problem

How We Work

• We optimise for deep work and uninterrupted time
• High ownership - you own it end to end
• Decisions backed by code and production outcomes

How to Apply

Show us AI systems you've shipped, evals you've built, and failures you've debugged in production. We don't care about years of experience or certifications. We care about what you've actually done.

About SocialMediaCheck

SocialMediaCheck is the automated intelligence layer for modern safeguarding. We build large-scale, real-time systems for safeguarding and risk detection. Our platforms operate under strict time, scale, and reliability constraints - where missing signals is not acceptable.