Sheraz Labs

Hi, I'm Sheraz.

I'm an AI engineer in London. I like sitting with the people who do the work, finding the part that eats their week, and shipping software that takes it off their hands, from the first conversation to a system running on AWS.

Right now that means agentic AI: systems where a model plans, calls tools and checks its own work, built with the unglamorous parts done properly, like logging every call, capping costs and deleting data on time. I hold an MSc in Data Science from Kingston and the AWS Solutions Architect certification.

How I got here

Shop floor, code, then AI.

I started behind the counter of an electronics store in Sialkot, working part-time while I studied Computer Science. Reading what customers actually bought, and restocking around it, took monthly sales from 50,000 to 180,000 PKR in two years. That's where I learned that data is only worth something when it changes a decision.

At university I led the team behind AssistivePro, a computer-vision project for accessibility that won government R&D funding. After graduating I worked as a Django developer, building a loan management platform from the first requirements meeting to a tool people used every day.

In 2024 I moved to London for an MSc in Data Science at Kingston. My placement year was at a facilities services company, where I audited 189 paper and WhatsApp processes, built the platform that replaced them, and learned how much of an operations team's week goes on work a machine should be doing.

Now I build agentic AI systems: agents that call real tools, keep track of state and cost, and run in production. The Video Editing Agent is my research playground for that, and Work Box is the app I ship and support on Google Play.

What I build

Three kinds of work, all shipped to production.

Agentic workflows

LangGraph agents that call real tools, keep state across a conversation, and account for every LLM call they make.

See the Video Editing Agent

AI-powered media & automation

Transcription, captioning and compositing pipelines - and the automation that replaces manual, repetitive operations work.

Try the research demo

Cloud deployment on AWS

Dockerised services behind nginx with automatic TLS, Postgres, persistent media volumes - designed and documented as C4 diagrams.

View the architecture

Toolkit

What I reach for, and use in production.

AI and machine learning

  • LangGraph
  • LangChain
  • Gemini
  • Groq
  • DeepSeek
  • Whisper
  • scikit-learn
  • TensorFlow
  • OpenCV

Backend and data

  • pandas
  • NumPy
  • Python
  • Django
  • PostgreSQL

Cloud and infrastructure

  • AWS
  • Docker
  • Nginx
  • Linux
  • Git

Experience and education

The dated version of the story above.

Experience

  1. Oct 2025 – Aug 2026

    Business Operations Analyst (placement)

    A facilities services company, London

    • Audited 189 operational forms across 5 departments, then scoped automation to the 82 that mattered most.
    • Built and deployed a Django platform on AWS (EC2, ECR, S3) that became the single source of truth for HR, control room and site reporting.
    • Cut weekly availability collation from 6 hours to 35 minutes, and turned a 3-hour incident report into one query.
    • Ran a build-vs-buy evaluation of local vision models, OCR libraries and commercial APIs that kept the platform on CPU-only compute.
  2. Dec 2023 – Sep 2024

    Junior Django Developer

    Steeline Corporation, Sialkot

    • Built a loan management platform in Django, from requirements with operations staff to a working tool.
    • Designed REST APIs connecting backend databases to reporting tools.
  3. Jan 2020 – Sep 2024

    Assistant Manager (part-time)

    Electronics store, Sialkot

    • Grew monthly sales from 50,000 to 180,000 PKR over two years by reading purchase data and reshaping stock.

Education and certifications

  1. 2024 – 2026

    MSc Data Science

    Kingston University London

    • Machine learning, deep learning, advanced analytics and data visualisation.
    • Guest speaker at the 2025 Placement Bootcamp, mentoring undergraduates.
  2. 2019 – 2023

    BS Computer Science

    University of Sialkot, Pakistan

    • Led the team behind AssistivePro, a computer-vision project that won PKR 70,000 in government R&D funding (IGNITE).
  3. May 2026

    AWS Solutions Architect – Associate

    Amazon Web Services (SAA-C03)

  4. 2022 – 2023

    Google Project Management and Data Analytics

    Professional certificates, Coursera

Working on an AI system? Let's talk.

Whether it's an agent that needs to reach production or a manual process that should be automated, tell me about it.