Skip to content

Zahoor Ahmed · Senior Software Engineer & Applied AI Engineer

I build production-ready AI systems for businesses.

I help startups and businesses build AI agents, RAG applications, LLM-powered products and workflow automations — from idea to production on AWS.

  • AI Agents
  • RAG
  • LLM Applications
  • MCP
  • Automation
  • AWS

Have an AI idea or a process you want to automate?

8+
Years Experience
6
Live Products
3
Countries Delivered
Zahoor Ahmed, AI engineer and senior software engineer

Services

Problems I solve

Clients don't buy technologies. They buy automated processes, working products and less manual work. Here is what that looks like.

AI Agents & Business Automation

Agents that call your APIs, work with your business data and carry a workflow through to the end — instead of a chatbot that only talks.

  • Tool-calling agents wired to your systems
  • Multi-step workflows with human approval where it matters
  • Guardrails, evaluation and logging from day one

RAG & Knowledge Assistants

Turn manuals, tickets, contracts and internal wikis into an assistant that answers from your own documents and shows its sources.

  • Document ingestion, chunking and embeddings pipeline
  • Hybrid retrieval with citations and access control
  • Answer-quality evaluation before and after launch

AI Product Development

Take an AI idea to a working SaaS or MVP — backend, frontend, database, payments and deployment, built to be extended rather than thrown away.

  • Scoped MVP with a clear release plan
  • Multi-tenant SaaS foundations, auth and billing
  • Production deployment on AWS with monitoring

AI Integration into Existing Software

Add OpenAI, Claude or self-hosted models, MCP servers and LLM features to the applications your team already runs — without a rewrite.

  • Model selection for cost, latency and data privacy
  • MCP servers and function-calling interfaces to your data
  • Fallbacks, rate limits and cost controls

Full-Stack & Cloud Engineering

The engineering around the model is what reaches users: APIs, front-ends, databases, Kubernetes on AWS and CI/CD pipelines that keep shipping.

  • TypeScript, Node.js, Python and Next.js / Angular front-ends
  • PostgreSQL, vector stores and search
  • Docker, Kubernetes, GitLab CI/CD and Grafana observability

Something else?

Not sure which of these you need?

Describe the workflow or the idea in a few lines. On a free 30-minute call I will tell you what I would build first, and whether AI is the right tool at all.

Book a discovery call

Case studies

Proof, not promises

Products I own and platforms I built for employers, each told as the problem, what I built and what it changed.

Screenshot of GuesstoLive

Multi-Tenant SaaS Platform · Own product — architecture to production

Guessto

Problem
Independent restaurants want their own website, digital menu, ordering and card payments, but cannot pay for a bespoke build and a developer per venue.
Solution
A multi-tenant SaaS: every venue gets its own subdomain, menu, allergen and diet filters, QR-code ordering, Stripe payments and transactional email — all served from one codebase and one deployment.
  • Next.js
  • TypeScript
  • PostgreSQL
  • Stripe
Read the case study
RankForge AI
In development

Autonomous SEO Engineering · Own product — in development

RankForge AI

Problem
SEO teams spend hours finding technical issues by hand, then write tickets and wait for developers — and rarely learn whether a fix moved anything.
Solution
An autonomous AI SEO engineer. It crawls the site, analyses pages with LLM agents, generates the code changes, opens them as GitHub pull requests for review and tracks rankings afterwards to close the loop.
  • LangChain
  • Playwright
  • Fastify
  • BullMQ
  • Prisma
Read the case study
Screenshot of RetromotionLive

E-Commerce Platform · Full-stack developer, Retromotion GmbH, 2022–2024

Retromotion

Problem
An auto-parts retailer selling through its own shop plus Amazon, eBay and Tyre24 needed one place to manage a large catalogue, search it fast and keep orders from every channel in sync without manual re-keying.
Solution
A microservices marketplace with Elasticsearch product search, a real-time admin dashboard over Socket.io, automated Amazon Seller, eBay and Tyre24 order synchronisation through SOAP and REST integrations, and bulk import/export workflows.
  • Vue.js
  • Node.js
  • TypeScript
  • Elasticsearch
Read the case study
Screenshot of TPConnectsLive

Airline Retailing · Software engineer, TPConnects Technologies, 2019–2022

TPConnects

Problem
Airlines and travel sellers needed an NDC-compliant retailing platform that could search, price and book across multiple GDS and airline suppliers, each with its own protocol and quirks.
Solution
Built the booking flow from scratch to production in Angular and React with NgRx and Redux managing complex booking state, and integrated GDS and airline suppliers through Java Spring Boot microservices on a fully AWS-native backend.
  • Angular
  • React
  • Java
  • AWS
Read the case study

More work

Screenshot of the MJ Carros Koblenz restaurant orderingLive

Built on Guessto · Restaurant Ordering

A live venue running on the Guessto platform — full digital menu, allergen and diet filters, and direct ordering on its own subdomain.

  • Next.js
  • Guessto
Screenshot of the DUDI sports community appLive

Sports Community App

Community platform for active lifestyles, shipping on iOS and Android alongside the web product.

  • React
  • Node.js
Screenshot of the UniKoop e-commerceLive

E-Commerce

Dutch home-shopping storefront with multi-category catalogue, promotions and a full checkout flow.

  • JavaScript
  • E-Commerce
DeutschFlow AIIn development

DeutschFlow AI

Own product · AI Language Learning

German language practice driven by conversational LLM agents and retrieval over graded source material.

  • RAG
  • LangGraph
  • OpenAI
VenueXplorerArchived

VenueXplorer

Venue Discovery

Venue discovery and booking experience built for the Singapore market.

  • React
  • Node.js

Why work with me

AI plus the engineering that ships it

  • AI plus serious software engineering

    The model is the easy part. The retrieval, the tool calling, the evaluation, the deployment and the monitoring around it are what reach users — and that is the work I have done for eight years.

  • Production experience, not demos

    Fleet telematics, an auto-parts marketplace and an IATA-certified airline platform: systems with real traffic, real money and real on-call.

  • Cloud and delivery included

    AWS, Kubernetes, GitLab CI/CD and Grafana come with the engagement. You get a system that is deployed and observable, not a notebook.

  • I ship my own products

    Guessto runs live restaurants today and RankForge AI opens pull requests on real repositories. I know what it takes to get from idea to paying users.

I bridge frontier AI models with robust, production-ready software architecture. Eight years of shipping fleet telematics, e-commerce and airline platforms taught me that the model is the easy part — the engineering around it is what reaches users.

Zahoor AhmedSenior Software Engineer · AI Engineer

How it works

From first call to production in four steps

  1. 01

    Discover

    A free 30-minute call about your business, your workflow and what you want to automate or build. You get an honest view of whether AI helps here.

  2. 02

    Architect

    Data, model choice, integration points and a delivery plan with a fixed first milestone you can hold me to.

  3. 03

    Build

    A working prototype first, then production code: tested, reviewed and shipped in small increments you can try each week.

  4. 04

    Launch

    Deployed on AWS with monitoring, evaluation and cost controls in place, and support after go-live.

Who I help

Built for teams without an AI department

Startups

Build your AI MVP without hiring a full engineering team, on foundations that survive your first customers.

Small & medium businesses

Automate repetitive workflows and internal processes: support triage, document handling, reporting, data entry.

Software companies

Add experienced AI engineering capacity to your team for a feature, a quarter or a product line.

Agencies

Deliver AI projects for your clients without building an AI team in-house. White-label welcome.

Pricing

Start small, scale when it works

Starting points, not quotes. Every engagement begins with the free discovery call, and fixed-price offers follow a written scope.

AI Discovery

€250fixed

Business and process analysis plus a technical AI roadmap you can act on, with or without me.

  • 90-minute workshop on your workflows and data
  • Written roadmap: what to automate first and how
  • Model, cost and data-privacy recommendation

Timeline: One week

Most chosen first

AI Prototype

from €1,500

A working AI proof-of-concept on your real data, so the decision to invest is made on evidence.

  • Agent or RAG prototype connected to your data
  • Evaluation set and quality report
  • Demo you can share internally

Timeline: 1–2 weeks

AI MVP

from €3,000

AI plus backend, frontend, database and deployment: a product your first users can log in to.

  • Scoped feature set and release plan
  • Auth, billing and multi-tenant foundations where needed
  • Deployed on AWS with monitoring and cost controls

Timeline: 4–8 weeks

AI Engineering

€70+per hour

Ongoing development, integration and technical support as part of your team.

  • Weekly or monthly capacity
  • Code review, architecture and pairing
  • Production support and incident response

Timeline: Flexible

FAQ

Questions clients ask before the first call

Anything missing? Ask it on the discovery call, or send it through the form below.

What kind of AI projects do you take on?

Agents that automate a business workflow, RAG assistants over company documents, LLM features inside existing software, and complete AI products or MVPs. If a task involves unstructured text, documents or decisions that currently need a person to read and route things, it is usually a good fit.

How long does an AI prototype take?

Typically one to two weeks. A prototype runs on your real data and comes with an evaluation set, so you see actual answer quality before committing to a full build.

Which models do you work with?

OpenAI, Anthropic Claude, Hugging Face models and self-hosted open-source models through Ollama. The choice follows your data-privacy requirements, latency and cost. EU-hosted and on-premise options are available when data cannot leave your infrastructure.

Can you integrate AI into the software we already have?

Yes. Most engagements add AI to existing systems through APIs, MCP servers or function calling rather than replacing them. I have worked with Node.js, Python, Java, Angular, React, Vue and PostgreSQL stacks, and with legacy SOAP and REST integrations.

What does it cost?

Discovery is €250 fixed, a prototype starts at €1,500, an MVP at €3,000, and ongoing engineering from €70 per hour. Every project starts with a free 30-minute call, and fixed-price work follows a written scope.

Where are you based, and do you work remotely?

I am based near Koblenz, Germany, and work remotely with clients across Germany, the EU and beyond in English. I have delivered production software in Germany, the UAE and Pakistan.

About

Eight years of production software

8+ years building production software across fleet telematics, e-commerce, airline technology and AI — in Germany, the UAE and Pakistan.

  1. Senior Software Engineer · PTC Telematik GmbH

    Dec 2024 – Present · Coblenz, DE

    Angular 19 / TypeScript PWA modules for fleet telematics, from TDD to deployment.

  2. Full Stack Developer · Retromotion GmbH

    Dec 2022 – Aug 2024 · Stuttgart, DE

    Microservices-based e-commerce platform for auto parts (Vue.js, Node.js, TypeScript, PostgreSQL).

  3. Software Engineer · TPConnects Technologies

    Oct 2019 – Oct 2022 · Dubai, UAE

    Flight booking platform (Angular, React, Node.js, PostgreSQL) from scratch to production.

Full CV (.docx)

5 roles in total · Mülheim-Kärlich, Germany · English and German

Education & certification

  • AI Engineering Programme

    Turing College · 2026

    Prompt engineering and LLM benchmarking, RAG with LangChain, function calling, and production AI agents with long-term memory.

  • Associate AI Engineer for Developers

    Datacamp · 2026

  • B.Eng. Software Engineering (BSCSE)

    Virtual University of Pakistan · 2014 – 2018

Technical stack

What it is built with

For the engineers on your side. Everything above is delivered with this stack, and the full list is in the CV.

  • TypeScript / JavaScript95%
  • Angular / NgRx95%
  • React / Next.js90%
  • Node.js / NestJS90%
  • Python / FastAPI88%
  • AWS / Kubernetes85%
  • AI / RAG / Vector DBs85%

Engineering Stack

  • TypeScript
  • React
  • Next.js
  • Angular
  • Node.js
  • Python
  • FastAPI
  • PostgreSQL
  • Neo4j
  • AWS
  • Docker
  • Kubernetes

AI Engineering

  • LangChain
  • LangGraph
  • OpenAI
  • Claude
  • Hugging Face
  • Chroma DB

Start here

Start with an AI discovery call

30 minutes · Free · Remote

Tell me about:

  • your business and who your customers are
  • the workflow that eats the most time today
  • what you would like to automate or build
  • your AI or product idea, however rough

I will tell you whether I think AI can realistically help, what I would build first, and what it would roughly take. No slides, no obligation.

Your details are used only to reply to you. No newsletter, no sharing.