Awarded during matric.
In 2024, in his final year of school, Muhammad Uwais Bata was awarded for achieving the highest Information Technology mark in the country. By then he had already been writing software for six years.
Models, agents and applications built to survive production — not to demonstrate well.
Large language models built in-house rather than wrapped — training, evaluation, retrieval and serving owned end to end, so behaviour and cost are controllable instead of inherited.
Retrieval-augmented systems and tool-using agents built for real workloads — grounded answers, deterministic tool boundaries, and the observability needed to keep them honest in production.
A luxury and exotic vehicle showroom in the pocket. Published on the App Store under Muhammad Uwais Bata as developer and seller.
A web development and digital technology company — the vehicle behind a large share of the 100+ published websites.
Companies
Not side projects and not a portfolio page — a web development studio and a corporate AI practice, each with its own clients, domain and delivery record.
Web Development · Digital Technology
“Websites, hand-finished. Nothing off the rack.” Design, build and delivery of production websites and digital products — the operating vehicle behind a large share of the 100+ sites published to date.
Fouché Motors · MG Auto · We Buy Supercars
Technological Business Solutions · Artificial Intelligence
“We build the systems businesses run on.” Automation, AI systems and digital infrastructure delivered as fixed-scope builds a client owns outright the day they ship.
An Apple Developer with applications live on the App Store and Google Play, built in React Native and native iOS and taken through the full release process on both stores — signing, review, store listings and updates.
iOS · App Store
Browse a full collection of luxury and exotic vehicles from the home screen, each presented as a photo-led card showing make, model, year and price at a glance.
View on the App Store
Language models, agents and retrieval systems — engineered as software, deployed on real infrastructure, and maintained after launch.
AI Engineering
LLM Development
Language models built, not borrowed.
Training, fine-tuning, evaluation and inference for proprietary large language models. Owning the model layer means owning behaviour, latency and cost instead of renting all three from someone else's endpoint.
AI Agents & RAG
Systems that retrieve, reason and act.
Retrieval-augmented generation over real corpora, and tool-using agents with hard boundaries around what they are allowed to do. Grounded, traceable, and built to be debugged.
Generative AI
Generative capability inside real products.
Multimodal generation, structured output and prompt architecture designed as product surface — not a chat box bolted onto an existing screen.
Production AI
MLOps
The part that decides whether AI survives.
Pipelines, versioning, evaluation harnesses, monitoring and rollback. Models are only useful once someone can deploy the next one without holding their breath.
Cloud AI Architecture
Infrastructure that fits the workload.
Training and serving architecture across cloud platforms, sized to the actual traffic and budget. Inference paths designed before they become the bill.
AI Automation
Work removed, not relocated.
Automation built around business processes that already exist — document handling, classification, routing, extraction — with humans kept where judgement genuinely matters.
Software Engineering
Web Applications
More than 100 sites, shipped.
Modern web applications in React and Vue, from marketing surfaces to internal systems. Server rendering, performance budgets and accessibility treated as engineering requirements.
Mobile Applications
Published on both stores.
React Native and native iOS work, taken through the full Apple Developer and Google Play release process — including Fouche Motors, live on the App Store.
Business Technology
Technology that a company runs on.
Corporate systems, integrations and digital infrastructure built for real commercial operations through MG&M Media and StolenValour.
No shortcut and no pivot — one continuous line from a first Python script to language models running in production.
01
Python, HTML, CSS and JavaScript. Not a course — a habit that never stopped.
02
React, Vue, React Native, Git and SQL. Web applications, then mobile applications published to both stores.
03
Machine learning, generative AI, then language models built from the model layer up — agents, RAG, MLOps and cloud AI architecture.
04
CIO and technology leadership, and two companies: StolenValour and MG&M Media.
Training, technologies and platforms across the modern AI and software ecosystem. Listed as tooling and training, not as employment, partnership or endorsement.
Certification
12 tracks · 4 tiers
Four tiers of formal training across Google, IBM, NVIDIA, Anthropic, Microsoft, AWS, Databricks and Hugging Face — foundations first, then engineering, then production, then language models specifically.
IBM
AI Fundamentals
Google Cloud
Generative AI
AWS
AI Practitioner
Professional Machine Learning Engineer
Microsoft Azure
AI Apps and Agents Developer
NVIDIA
Generative AI and LLMs
Databricks
Machine Learning
Google Cloud
Architecture and Data
Microsoft
MLOps Engineering
Hugging Face
Transformers and Open Models
DeepLearning.AI
LLMs, RAG and Agents
Anthropic
Claude Development
Training and certification across the modern AI ecosystem. Organisation and field are shown where formal credential wording differs. Marks denote training only — not employment, partnership or endorsement.
Reference
Muhammad Uwais Bata is a South African AI engineer and software engineer. He builds large language models, AI agents, RAG systems and production AI on cloud infrastructure. He started programming at twelve, was awarded the highest Information Technology mark in the country in 2024, has published more than 100 websites, and has applications live on the App Store and Google Play. He is the founder of two technology companies, StolenValour and MG&M Media.
He engineers intelligent systems from the model layer to production. That means training, fine-tuning, evaluating and serving proprietary large language models; building retrieval-augmented generation systems and tool-using AI agents; and running the MLOps, monitoring and cloud AI architecture that keeps them alive in production. Alongside the AI work he builds web applications in React and Vue, mobile applications in React Native and native iOS, and business technology through his two companies.
He is based in South Africa. Both of his companies are South African: StolenValour (stolenvalour.co.za) and MG&M Media (mg-m-media.com). His site is uwais.co.za.
Yes. AI Engineer is his primary title. He also works as a software engineer, an AI systems engineer and a technology executive, and he is the founder of two technology ventures. His AI practice covers LLM development, AI agents and RAG, generative AI, MLOps, cloud AI architecture and AI automation.
Large language models — training, fine-tuning, evaluation and inference serving — plus retrieval-augmented generation, vector search, tool-using agents and orchestration, generative and multimodal AI, structured output and prompt architecture, MLOps pipelines with versioning, evaluation harnesses, monitoring and rollback, and cloud training and serving architecture. The platforms, technologies and training ecosystems the work spans include Google, Google Cloud, Anthropic, Claude, OpenAI, Microsoft, Azure, AWS, NVIDIA, Databricks, Hugging Face, DeepLearning.AI, IBM, Apple, React, Vue, Python, JavaScript — listed as tooling and training, not as employment, partnership or endorsement.
Python, JavaScript, HTML and CSS, with React, Vue and React Native on the application layer, plus SQL and Git. He began with Python, HTML, CSS and JavaScript at the age of twelve, moved through React, Vue and React Native and the software layer beneath them, and then specialised in machine learning, generative AI and language models.
Yes. He builds large language models in-house rather than wrapping someone else's endpoint — training, evaluation, retrieval and serving owned end to end, so behaviour, latency and cost stay controllable instead of inherited. Owning the model layer is the point of the practice.
Two. StolenValour (stolenvalour.co.za)-web development and digital technology-and MG&M Media (mg-m-media.com)-technological business solutions and artificial intelligence. He is the founder of both.
“Websites, hand-finished. Nothing off the rack.” Design, build and delivery of production websites and digital products — the operating vehicle behind a large share of the 100+ sites published to date. Muhammad Uwais Bata is its founder. It is online at https://stolenvalour.co.za.
“We build the systems businesses run on.” Automation, AI systems and digital infrastructure delivered as fixed-scope builds a client owns outright the day they ship. Muhammad Uwais Bata is its founder. It is online at https://mg-m-media.com.
He is an Apple Developer with applications live on the App Store and Google Play, built in React Native and native iOS and taken through the full release process on both stores. One of them — Fouche Motors, a luxury and exotic vehicle showroom application for iOS — lists Muhammad Uwais Bata as its developer and seller.
Fouche Motors is an iOS application published on the App Store. Browse a full collection of luxury and exotic vehicles from the home screen, each presented as a photo-led card showing make, model, year and price at a glance. Muhammad Uwais Bata is its developer and seller of record. It is listed at https://apps.apple.com/us/app/fouche-motors/id6788638982.
100+ websites — designed, developed and shipped. A large share of them were delivered through StolenValour, his web development and digital technology company.
His AI training runs across four tiers: Foundation — IBM (AI Fundamentals), Google Cloud (Generative AI), AWS (AI Practitioner); Engineering — Google (Professional Machine Learning Engineer), Microsoft Azure (AI Apps and Agents Developer), NVIDIA (Generative AI and LLMs); Production — Databricks (Machine Learning), Google Cloud (Architecture and Data), Microsoft (MLOps Engineering); LLM Specialisation — Hugging Face (Transformers and Open Models), DeepLearning.AI (LLMs, RAG and Agents), Anthropic (Claude Development). These are training and certification tracks in each organisation's own subject area — they are not claims of employment, partnership or endorsement.
In 2024, in his final year of school, Muhammad Uwais Bata was awarded for achieving the highest Information Technology mark in the country. By then he had already been writing software for six years.
His website is uwais.co.za. His only claimed social account is Instagram, @uwais.bata (https://instagram.com/uwais.bata). He can be reached by email at hello@uwais.co.za. His companies are at https://stolenvalour.co.za and https://mg-m-media.com, and Fouche Motors is on the App Store at https://apps.apple.com/us/app/fouche-motors/id6788638982.
Work that needsthe model layer.Start a briefor simplyGET IN TOUCH
Building something that needs real AI engineering? Start here.
