Agentic AI · LLM apps · RAG

Your agent
demo worked.
Now let’s make it
finish the job.

Every time, on budget. I’m Saqib: I build LLM, RAG and agent systems that survive real users, real data and real bills.

Studio portrait of Saqib Ejaz in a charcoal blazer
Scroll to exploreFrom prototype to dependable product
01 / Where I come in

When “it works”
isn’t working.

You’ve built something. Now it has to survive actual users, actual data and actual bills.

01 / AUTOMATION+

“The agent stops halfway.”

Workflows that recover, retry and actually finish.

02 / RELIABILITY+

“It sounds sure. It’s wrong.”

Fix retrieval, add evaluation, earn user trust.

03 / COST+

“Our AI bill keeps climbing.”

The right model and architecture for the job, not the biggest one.

04 / MODEL CHURN+

“Should we switch models again?”

Test what improves outcomes before changing stacks.

05 / ACCESS+

“Who can see our data?”

Access, keys and data boundaries made explicit.

02 / Work behind the words

AI that shipped,
not slideware.

25%

operational efficiency gain from LLM automation at WeMove

5,000+

moves supported by the WeMove platform

8+

years shipping production software, end to end

WeMoveUS · remote
Team lead

AI-powered scheduling and logistics microservices in NestJS and PostgreSQL. LLM automation that lifted operational efficiency 25% on a platform supporting 5,000+ moves.

Current role2025 – now
Lead / GenAI engineer

Distributed systems and high-performance APIs for large-scale code analysis and dependency mapping. LLM + retrieval automation pipelines. AWS and Azure with Docker and Terraform.

Amperor Strata ↗PyPI · amperor-strata
Creator

Runs Mixture-of-Experts models bigger than your RAM on Apple Silicon by paging experts from SSD. OpenAI-compatible server. Published on PyPI.

AlayaCareCanada
Tech lead · secure messaging

Led secure messaging: OAuth 2.0 / JWT and the serious version of “who can see what”.

Selected outcomes from previous roles.
03 / Stack

Tools, not religion.

Whatever gets it right, fast and affordable.

OpenAIAnthropicRAG pipelinesAgentic workflowsVector databasesMoE inference (Strata)PythonFastAPINode.jsNestJSPostgreSQLRedisAWSAzureDockerTerraform
04 / Pick your flavour

Same engineer.
Your problem.

Choose the one that sounds like yours. Each page is the short version for that job.

05 / Let’s talk

Tell me what’s breaking.

No 50-slide strategy deck. A direct conversation about the problem and what it takes to fix it.