03 AI Engineer Remote / Islamabad

Robotech Solution · Islamabad, Islamabad Capital Territory, Pakistan

Ship LLM apps with real users - not demos.

You'll build RAG systems, eval harnesses, and agent workflows for clients across health, finance, and enterprise. We measure shipped accuracy and $/request, not benchmark scores.

Stack

Python FastAPI pgvector Anthropic / OpenAI vLLM Modal

What you'll do

• 01 Build retrieval + reranking pipelines
• 02 Design eval harnesses with the client's domain experts
• 03 Operate prompt + model rollouts behind flags
• 04 Track cost and accuracy in dashboards

What you need

Have shipped at least one LLM product to real users

Strong Python and at least one of: FastAPI, vLLM, LangChain

Understand retrieval, chunking, and reranking deeply

Comfortable instrumenting cost

Nice to have

ML platform background

Hiring process

• 01 30-min intro call
• 02 Eval-design conversation (60 min)
• 03 Take-home (4 hours, paid)
• 04 Team meeting + offer