name: Zainab Hamid
role: AI Engineer @ Liztek
focus: Agentic systems, Gen AI, Python backends
based_in: Lahore, Pakistan 🇵🇰
fueled_by: cold coffee + a stubborn debugger
mood: "let's just ship it and iterate"

I take messy real-world data, weird stakeholder requirements, and LLM behavior that nobody fully understands — and turn it into backends that don't fall over on a Tuesday afternoon.

Currently: building multi-agent data quality systems for an enterprise financial product piloting at a top-tier Pakistani bank.

_{me, mid-debug, lying to myself
about "one more fix"}

_{Day job · Liztek}

Agentic data quality platform for enterprise finance. Validator → Parser → Anomaly Detection → Resolution agents, all coordinated, all logged, all auditable.

_{LangGraph · FastAPI · PostgreSQL}

_{Side build}

WhatsApp-native AI agent that parses Urdu / Hinglish RSVPs (yes, including "Inshallah" and "kal batata hoon") into clean state.

_{LLMs · WhatsApp API · Cultural NLU}

_Personal

Vector search + RAG pipeline that actually returns relevant chunks, not just technically relevant chunks.

_{Pinecone · LangGraph · Embeddings}

"The most dangerous phrase in software is 'it works on my machine.' The second most dangerous is 'the LLM said it would do that.'"

I genuinely believe a good print() statement has resolved more existential crises than any therapist could. The hierarchy of debugging:

print() → pdb → re-read the docs → ask the LLM → cry
 ↑
 (this step is non-optional)

Three things I'll always pick:

• ☕ Cold coffee over hot tea — fight me
• 🐍 Python over almost anything (sorry JavaScript, you tried)
• 🎯 Shipping a 70% solution today over a 100% solution in three months