Skip to content
Razi Ullah.
AI Solutions Engineer

Production AI that cuts cost 60–80% and scales to real-world constraints.

Building production AI systems, intelligent applications, and scalable platforms.

View ProjectsContact→

A cel-shaded graphics card: shroud, fans, heatsink fins, GPU die, VRAM, PCB, backplate, PCIe connector and I/O bracket.

At a glance

Reasons teams work with Me

Cost-efficient by design

On-demand GPU orchestration, distributed inference, and lean pipelines that cut compute spend 60–80% without sacrificing latency.

  • GPU orchestration & async processing
  • Distributed inference
  • Real-time deployment

Production-grade, not prototypes

Systems built for live users, strict SLAs, and 50TB+ scale — reliability and performance that hold up under real load.

  • 100+ users, 50TB+ data
  • Vision, LLMs & agents
  • Benchmarked & monitored

End-to-end ownership

From data pipelines and fine-tuning to backend APIs and deployment — one engineer accountable for the whole system.

  • Data → model → deployment
  • FastAPI backends & RAG
  • Automation that scales 5–10×

The impact

$100K+/year.

Cutting infrastructure cost while scaling AI systems to real-world constraints — saved via ML automation, 50TB+ data, and 100+ users.

Cost-efficientScalableProduction-readyMeasurable impact
10+
Projects Delivered
10+
Satisfied Clients
5+
Industries Served
100%
Client Satisfaction Rate

About

Proof, not promises

HELLOmy name isRazi

I started in production customer operations, working with live systems, strict SLAs, and real users—where reliability and performance directly impact business outcomes. This foundation shaped my approach to engineering: building systems that don't just work in theory, but operate reliably in production. I transitioned into software engineering and ML systems, working across data pipelines, model integration, and backend infrastructure—gaining end-to-end ownership of AI systems. Today, I focus on building production-grade AI systems across vision, LLMs, and infrastructure, with an emphasis on cost efficiency, scalability, and real-world impact—including systems that deliver 5–10× efficiency gains and 60–80% cost reductions.

Tech Stack

The stack behind the work

The tools I reach for across computer vision, LLM systems, and the backends that ship them to production.

  • Python
  • JavaScript
  • React
  • FastAPI
  • REST APIs (JWT/OAuth)
  • Async Processing
  • PostgreSQL
  • Redis
  • pgvector
  • Pinecone
  • GCP (Compute Engine, Cloud Storage)
  • Docker
  • CI/CD
  • LLM Systems (RAG, Agents, Tool Calling, Evaluation)
  • Fine-tuning (SFT, PEFT, RLHF)
  • Domain Adaptation
  • Prompt Engineering
  • Computer Vision
  • NumPy
  • Pandas
  • LangChain
  • LangGraph
  • Hugging Face
  • TensorFlow
  • Scikit-learn
  • Voice AI (STT/TTS)
  • ElevenLabs
  • Git
  • GitHub

Solutions

What I build

End-to-end AI systems — from data pipelines to production deployment.

AI Systems & Agents (LLMs, RAG, Automation)

Design and deploy production-ready AI agents and LLM systems integrated into real workflows, reducing manual effort and operational overhead.

Computer Vision & Data Pipelines

Build scalable vision systems and data pipelines (annotation, labeling, model integration) handling large-scale datasets and real-time constraints.

ML Infrastructure & Distributed Systems

Develop cost-efficient ML infrastructure (GPU orchestration, async processing, pipelines) to reduce compute costs and improve scalability.

Workflow & Business Automation

Replace manual processes with reliable automation systems, improving efficiency by up to 5–10×.

Data Annotation & Dataset Engineering

Design and scale annotation systems, quality pipelines, and dataset workflows for training high-performance models.

Custom Backend & AI-Integrated Systems

Build robust backend systems (APIs, orchestration layers, integrations) powering AI-driven applications in production.

Selected work

Projects & delivery

View all projects

Testimonials

What clients say

Feedback from clients and collaborators on AI, automation, and production systems.

Working with Taha A has been an outstanding experience from start to finish. His professionalism, attention to detail, and deep understanding of the task requirements truly set him apart. He delivered high-quality work on time and went above and beyond to ensure everything met (and exceeded) expectations. Taha is an excellent communicator — always responsive, clear, and open to feedback. He demonstrated a strong work ethic and a genuine commitment to achieving the best possible outcome. I would highly recommend Taha to anyone looking for a reliable, skilled, and proactive freelancer. Looking forward to working with him again on future projects!
OH
Osheen H.
Client · Google App Script
Taha is highly professional and reliable. His proactive 'can-do' attitude and commitment to timely delivery are truly commendable. He always goes one step ahead to complete the assigned tasks. He's an invaluable asset for any organization.
TS
Talha S.
Team Lead · Customer Services
Taha worked with us on automating our appointment scheduling workflow, delivering a complete solution deployed as a Chrome extension. He handled both the development of the automation script and integration with our existing extension seamlessly. The system significantly reduced manual effort and improved operational efficiency. Taha demonstrated strong technical expertise, professionalism, and attention to detail throughout the project. He was reliable, proactive, and committed to delivering a high-quality production-ready solution. I highly recommend him for AI automation, browser automation, and software development projects.
AZ
Asadullah Z.
Client · Chrome Extension for Automated Appointment Scheduling
Painted as you scrollFresh off the wall
Let's build

Move beyond the hype.
Ship real AI.

Whether you need an AI system built, a pipeline scaled, or a workflow automated — let's talk about what you're building.

Start a conversation →Explore solutions