Here is what you'll actually do.

You'll bridge the gap between business problems and AI implementation. For 1-2 concurrent projects, you will digest client requirements, design systems involving LLMs and vector databases, rapidly prototype the solution, build the production pipeline, and hand off an intelligent API or agent system. You ship working AI, not just Jupyter notebooks.

Recent task cards you would have pulled.

[ RAG Systems ]Complex conversational interfaces with semantic search
[ Agentic Workflows ]Multi-agent frameworks, LangChain pipelines, automated task execution
[ Python APIs ]FastAPI servers wrapping model calls with low latency

We are extremely specific about who fits.

You're right for this role if...

You've shipped something real and can show us the commit history
You understand that a reliable 80% solution beats a theoretical 99% accuracy model that never ships
You communicate progress before you're asked for an update
You follow the AI space hourly and know which new model invalidates our current approach

This role isn't right for you if...

You need detailed tickets with every edge case spelled out
You measure success by hours logged, not features shipped
You're looking for a stepping stone — we want people who want to stay

Tools of the trade.

Must Have:
PythonOpenAI API / ClaudeLangChain / LlamaIndexVector DBs (Pinecone/Weaviate)FastAPI
Nice to Have:
Hugging FaceAWS/GCP ml-ops deploymentsNext.js for quick UIs
We'll teach you:
Advanced prompt engineering for specific edge casesTypescript backend basics

Honest numbers. No fluff.

Base Salary
PKR 1M - 1.5M. Determined by your past shipped work.
Performance Track
USD Performance Track after 6 months. For top performers with high sprint velocity.
Absolute Flexibility
Async, outcome-based. No time tracking. Deliver features, not hours.
Public Growth
Case study credit. Your work becomes your portfolio. Your name goes on it.

How we decide.

01

Apply

~10 min

Form below. No CV required.

02

Skill Task

48 hrs

You'll receive a dataset and a problem statement. Build a simple RAG pipeline using a vector DB and an LLM to answer questions accurately with citations. Return a working API endpoint. 48 hours.

03

30-min Call

~30 min

We discuss your code.

04

Offer

24 hrs

Yes or No, within 24 hours.

Build with us.

Submit your best work to start the process.

We don't review CVs. We review work.