AI Engineer · Computer Vision · Generative AI

I build AI that
sees, reasons
& ships.

I'm Ata Ul Haq—also known online as Ata Satti—an AI engineer who turns ambitious ideas into reliable products, connecting models, APIs, and thoughtful interfaces into systems people can actually use.

Ata Ul Haq
AI systemsModel → Product
Focus / 01Multimodal intelligence
Focus / 02Generative products
Computer VisionGenerative AILLM ApplicationsFull-Stack ProductModel Orchestration

Selected work · 2024–26

Systems with
something at stake.

Multimodal moderation, creative tooling, medical vision, and 3D generation—built beyond the prototype.

Also shipped

Full-stack delivery and focused vision systems.
Orvia Studios landing page05

Full-stack SaaS

Orvia Studios

A live agency platform with a marketing site, role-based client portal, admin operations, analytics, project tracking, and invoicing.

Next.js · MongoDB · NextAuth · Recharts · Vercel

Oscillation Records logo06

Full-stack music platform · Live

Oscillation Records

A production platform for a Manchester independent record label, unifying artists, releases, streaming, submissions, accounts, analytics, and a full content-management backend.

Next.js 15 · TypeScript · Prisma · MongoDB · NextAuth · AWS S3

Made using Claude + CodexVisit live site

Earlier work · Selected

The foundations.

Four early projects that shaped the range behind today's end-to-end AI systems—from vision and analytics to security and full-stack delivery.

Capabilities

From raw input
to useful output.

The strongest AI work is not just a model. It is the orchestration, product judgment, and engineering around it.

01

Vision & perception

Detection, classification, segmentation, OCR, frame sampling, and production inference across medical and moderation use cases.

  • PyTorch
  • TensorFlow
  • OpenCV
  • YOLO v8/v11
  • U-Net
  • EfficientNet
02

Generative systems

Multi-provider image workflows, LLM applications, RAG, prompt systems, video, and 3D-generation orchestration.

  • OpenAI
  • Gemini
  • FLUX
  • Stable Diffusion
  • RAG
  • Hunyuan3D-2
03

Product engineering

The application layer around the model: resilient APIs, polished interfaces, authentication, databases, billing, and deployment.

  • FastAPI
  • Next.js
  • React
  • Three.js
  • MongoDB
  • Docker

How I work

Model to product,
without the handoff gap.

01

Frame the system

Start with the user decision, then define modalities, constraints, evaluation signals, and the right model mix.

02

Engineer the path

Build observable pipelines with resilient APIs, sensible timeouts, guardrails, and swappable providers.

03

Ship the experience

Turn the pipeline into a fast, clear interface and take it through deployment—not just a notebook demo.

About Ata Ul Haq

Ata Satti online · AI Engineer / Product builder

I like the hard part between“the model works” and“the product works.”

My work spans computer vision, generative AI, and LLM applications. I build across the stack—from preprocessing and inference to backend architecture and polished frontends.

I am especially drawn to multimodal systems: products where different models, data types, and user workflows have to behave like one coherent experience.

4Modalities handled in one moderation system
3Swappable 3D generation backends
2Medical imaging pipelines in Dental AI
1Engineer across model, backend, and interface

Education

Built on
data science.

Feb 2022—Feb 2026

Bachelor's degree

Data Science

COMSATS University IslamabadCGPA 3.65 / 4.0
Aug 2018—Sep 2020

Higher secondary

HSSC

F.G. Sir Syed College, RawalpindiGrade A+
Mar 2016—Aug 2018

Secondary school

SSC

APSACS (FWO), RawalpindiGrade A+

Start a conversation

Have an ambitious
AI problem? Let's build.

atasatti20@gmail.com