10 years in the field
I have spent a decade taking ideas the rest of the way: into models that can be checked, Python platforms people rely on, TypeScript products teams enjoy opening, and CI/CD pipelines that ship on a Thursday without a war room. AI, machine learning, Django, FastAPI, and delivery live in the same practice — not in separate slides.
Python products, from a first prototype to software a team could run every day.
Django platforms with real permissions, real data, and users who stayed.
Models leaving the notebook and landing behind production APIs.
AI systems wrapped in TypeScript products and releases that can roll back.
Skills
Depth across AI and machine learning, Python backends, TypeScript interfaces, and the DevOps practice that gets a change into production safely.
Models that leave the notebook: trained, evaluated, grounded, and served where a product can call them.
APIs and platforms with contracts, tests, and a database story that holds up.
A change should be built, tested, promoted, and reversible. That is the job.
Typed interfaces in front of the APIs, so the product and the backend stay in agreement.
Recent projects
AI and machine learning, Python, Django, FastAPI, TypeScript, and the pipelines around them. Filter by the layer you care about.
12 projects
AI · FastAPI · TypeScript
A private-document assistant. Uploads are chunked, embedded, and answered with the page they came from. FastAPI serves the retrieval, and a TypeScript workspace is where the team actually asks the question.
AI / ML · Python
Demand studio for operations. Python models forecast weekly volume, flag drift, and publish an accuracy note beside every run.
Computer vision · Python
Inspection models watch a production line. Frames are scored in Python, and a person confirms a defect before it leaves the station.
FastAPI · Python
A FastAPI service that turns raw features into scored signals. Pydantic keeps every response typed for the apps that consume it.
Django · Python
Multi-tenant CRM on Django and Django REST Framework. Pipelines, permissions, and activity history live in one place the team actually opens.
Django · Python
A hiring desk in Django. Roles, scorecards, and interview loops stay with the candidate, so nobody is keeping the real process in a side sheet.
DevOps · CI/CD
Delivery control for services on Kubernetes. Pull requests run GitHub Actions, build images, and stop at a human gate before production.
CI/CD · DevOps
Shared CI templates for a microservice fleet. Each repo inherits test, scan, and canary steps, and a failed gate stops the release.
TypeScript
A TypeScript analytics board. Funnels, cohorts, and experiment reads render from one typed client instead of a pile of one-off queries.
TypeScript · FastAPI · AI
A team workspace with a model in the room. The TypeScript UI streams from a FastAPI backend that keeps prompts, tools, and transcripts on the server.
MLOps · Python · CI/CD
Training jobs with a paper trail. Python runners fit a model, log the metrics, register the artifact, and only then let CI promote it.
FastAPI · TypeScript
Usage billing as an API. FastAPI meters events and issues invoices, while a TypeScript console lets finance review a customer without touching the database.
Testimonials
Notes from teams who asked for an AI system, a Python platform, or a release process they could trust after the kickoff call was over.
We had years of documents and a vague plan to “add AI.” Waqas shipped AetherRAG as a product: a FastAPI service, a TypeScript workspace, and answers that show the page they came from. The team stopped guessing which chatbot was right.
AetherRAG · AIThe forecast lived in a notebook only I could run. ForgeCast made it a Python service with a schedule, a drift check, and a report operations can read without calling me.
ForgeCast · MLHe treated delivery as part of the product. Tests, image builds, a staging promote, and a production gate. Our release channel went quiet, which is the compliment.
Harbor Deploy · CI/CDTalentGrid is the first hiring system the team did not work around. Django permissions match how we actually hire, and a candidate no longer disappears between a sheet and an inbox.
TalentGrid · DjangoThe billing API and the console agree with each other. FastAPI on one side, TypeScript on the other, and a contract finance can trust. That is harder than it looks, and he did not skip it.
Ledgerline · FastAPI