Open to opportunities

Hi, I'm
Hanzala Kashif

Most AI demos die in staging. Mine run in production, serving millions of users, processing millions of records, and earning revenue.

View My Work Download CV
Portrait photo of Hanzala Kashif

Who I am

Full Stack AI Engineer

Hanzala Kashif, AI Engineer based in Pakistan

I'm a Full Stack AI Engineer with 5+ years building enterprise-grade AI systems. Not Jupyter notebooks. Not proof-of-concepts. Real pipelines, real users, real outcomes.

I own the full path: data ingestion, RAG retrieval, multi-agent orchestration, and production deployment, with the reliability, observability, and compliance posture enterprise systems actually demand.

I lead AI and backend engineering for production systems and I'm open to Senior AI Engineer, AI Architect, and founding engineer roles. Open to relocation: UAE, GCC, Europe, Malaysia.

5+ Years Experience
Millions Users Served
89% Upwork JSS

What I've shipped

Projects

RAG

Graph & Contextual RAG on Azure AI Search

Designed Graph and Contextual RAG on Azure AI Search delivering 35% higher retrieval groundedness vs baseline vector search, with p95 query latency under 800ms across a 100,000+ document corpus at enterprise scale.

LangChain Azure AI Search Python Hybrid Search
View on GitHub →
Multi-Agent

LangGraph Multi-Agent Orchestration System

Built autonomous decision pipelines using LangGraph running concurrent production workloads, fully traced in LangSmith. Standardized reusable agent templates that cut average feature delivery time by 40%.

LangGraph LangSmith FastAPI Azure
View on GitHub →
Observability

LangSmith AI Observability Pipeline

Built a LangSmith observability pipeline with continuous groundedness monitoring and automated alerting that cut production silent-failure rate by 50%, giving engineering teams full visibility into LLM behavior in production.

LangSmith Python Automated Alerting
View on GitHub →
ETL

Async Batch ETL Across 5 Heterogeneous Sources

Designed and shipped an async batch ETL pipeline across 5+ heterogeneous data sources with intelligent job scheduling, delivering 60% lower data refresh latency and enabling real-time analytics on enterprise datasets.

Python FastAPI PostgreSQL Redis
View on GitHub →
Backend

Zero-Downtime B2B Platform Merge

Led a zero-downtime migration merging 2 live B2B platforms into a single Django system serving millions of active users. Delivered on schedule with zero production incidents over a 12-month engagement.

Django PostgreSQL Redis CI/CD
View on GitHub →
Caching

Redis Caching Layer for LLM Inference

Built a Redis caching layer that delivered 40% faster API responses and eliminated redundant LLM inference calls, directly lowering monthly inference spend for a high-traffic production system.

Redis FastAPI Python LangChain
View on GitHub →
Real-Time

Real-Time Social & Networking Apps

Built 3 real-time social and networking apps on Django REST featuring WebSocket chat, Agora video calling, push notifications, and activity feeds serving concurrent live users. Added one-tap sign-in via JWT and OAuth2 across 4 providers.

Django WebSockets Agora JWT OAuth2
View on GitHub →
Automation

Factory Production Modeling System

Modeled multi-stage factory production logic in Django covering takt time, machine routing, buffers, and assembly workflows across 5 product configurations. Eliminated 8+ hours per week of manual reporting with automated PDF and Excel analytics exports.

Django Python PDF Export Excel
View on GitHub →

Tools & technologies

Tech Stack

AI & Machine Learning

LangGraph LangChain OpenAI RAG Multi-Agent Systems LangSmith Prompt Engineering Hybrid Search

Backend Development

Python FastAPI Django Microservices WebSockets REST APIs

Cloud & DevOps

Azure Functions Azure Cosmos DB Azure AI Search AWS Docker Kubernetes Terraform CI/CD

Databases & Data

PostgreSQL Cosmos DB Redis DynamoDB

Standards & Compliance

OAuth2 GDPR SOC2-aligned System Architecture

Career history

Experience

Team Lead, AI & Backend Engineer Horizon Tech
April 2026 - Present · Karachi
  • Set the AI architecture standard for the team: reusable RAG, agent-orchestration, and cloud-native backend patterns now used as the default across new client builds
  • Own delivery end to end from requirements to production, holding a zero-downtime release record across the systems I ship
  • Driving the roadmap for next-generation multi-agent systems mapped directly to enterprise client requirements
Freelance AI & Backend Engineer Upwork
January 2023 - Present · Pakistan
  • Delivered production RAG pipelines on LangGraph, LangChain, and Azure AI Search for global clients, holding p95 latency under 800ms on enterprise-scale corpora
  • Maintained an 89% Job Success Score across engagements through end-to-end ownership and production-grade delivery
Senior Backend AI Engineer Fleekbiz Pakistan
September 2025 - April 2026 · Karachi
  • Shipped 3 production AI systems end to end and cut average feature delivery time 40% by standardizing reusable agent templates and infrastructure
  • Designed Graph and Contextual RAG on Azure AI Search: 35% higher retrieval groundedness vs baseline vector search, p95 query latency under 800ms at enterprise scale
  • Cut production silent-failure rate 50% by building a LangSmith observability pipeline with continuous groundedness monitoring and automated alerting
Backend AI Engineer Fleekbiz Pakistan
September 2024 - September 2025 · Karachi
  • Built a RAG system over 100,000+ documents on LangChain plus Azure AI Search with sub-second response, deployed to a live enterprise client base
  • Cut data refresh latency 60% across 5 heterogeneous sources via async batch ETL and intelligent job scheduling
  • Reduced average API response time 40% and eliminated redundant LLM inference calls with a Redis caching layer, directly lowering monthly inference spend
Python Engineer Brandigo
January 2024 - December 2024 · Pakistan
  • Led a zero-downtime migration merging 2 live B2B platforms into one Django system serving millions of users, delivered on schedule with zero production incidents over a 12-month engagement
  • Reverse-engineered a large undocumented Django codebase with no handover and shipped production features inside the first week
  • Cut PostgreSQL load on hot-path queries 35% with targeted Redis caching, improving response times on high-frequency endpoints
Python Developer Crystallite Pakistan
January 2023 - February 2024 · Pakistan
  • Built 3 real-time social and networking apps on Django REST: WebSocket chat, Agora video, push notifications, and feeds serving concurrent live users
  • Cut registration drop-off by adding one-tap sign-in: JWT plus OAuth2 across 4 providers (Google, Apple, Facebook, LinkedIn)
Python Developer 3Const
December 2021 - January 2023 · Pakistan
  • Modeled multi-stage factory production logic in Django (takt time, machine routing, buffers, assembly workflows) across 5 product configurations
  • Eliminated 8+ hours per week of manual reporting with an automated PDF and Excel analytics export system

Background

Education & Certifications

Bachelor of Science, Computer Software Engineering
Virtual University of Pakistan
June 2023 - August 2027
Computer Software Engineer, Computer Science
Aptech Pakistan
October 2021 - October 2024

AWS Educate Certifications

AWS AWS Educate Getting Started with Databases
AWS AWS Educate Getting Started with Serverless
AWS AWS Educate Introduction to Generative AI
AWS AWS Educate Machine Learning Foundations
AWS AWS Educate Getting Started with Storage

Get in touch

Let's work together

I'm open to Senior AI Engineer, AI Architect, and founding engineer roles. Available for relocation to UAE, GCC, Europe, and Malaysia. I reply within 24 hours.

Email LinkedIn GitHub

Open to: Senior AI Engineer · AI Architect · Founding Engineer