AI & machine learning development
AI and machine learning, built into your product.
I build AI that does real work in production: assistants connected to live business data, AI receptionists and calling agents, document processing and machine learning models trained on your data. The focus is always the same: reliable results and safe behaviour, not just an impressive demo.
- 7+ years in production
- Remote · clients on 4 continents
- Fixed price or monthly retainer
What I can build for you
LLM-powered apps
Assistants and features built on GPT, Claude and open-source models, tuned for your use case.
AI agents
Agents that use your app's tools to complete tasks, with permissions, confirmations and logs.
Chat with your documents (RAG)
Search and answer questions over your files, knowledge base or database.
Voice AI
AI receptionists and calling agents that answer, qualify and book.
Custom machine learning models
Prediction, classification and scoring models trained and deployed on your data.
AI strategy and prototypes
Find where AI saves the most time in your business, then prove it with a working prototype.
Stack: Python · scikit-learn · OpenAI · Anthropic Claude · Embeddings and vector search · MCP · Laravel · FastAPI
Read how I built it: How I Connected ChatGPT and Claude to a Laravel SaaS (Safely) with MCP
Related work
QistPay × ChatGPT & Claude
An MCP connector that lets QistPay owners query and update their business data from ChatGPT and Claude, with OAuth, per-action permission checks, confirmations and audit logs.
Read the case study →AI Canadian Solutions
AI receptionists, AI calling agents, CRM automation and virtual assistants for Canadian businesses, with a React front end.
Visit aicanadiansolutions.ca →How we work together
- Intro call. A free 30-minute call to understand your goals, users and timeline.
- Clear proposal. Scope, milestones and a fixed price or monthly plan, in writing.
- Build in the open. Regular demos and progress updates; you see working software early.
- Launch and support. Deployment, monitoring and ongoing improvements after go-live.
Common questions
What kind of AI can you add to my business?
Common projects are assistants that answer questions from your data, automatic document processing, AI phone receptionists, and models that predict things like demand, risk or churn.
Do you train custom machine learning models?
Yes. When an off-the-shelf LLM isn't the right fit, I train and deploy models on your own data for prediction and classification tasks.
How do you keep AI from making costly mistakes?
By design: limited permissions, checks on every action, previews and confirmations before changes, and logs of everything the AI does.
How long does an AI prototype take?
A focused prototype usually takes a few weeks. It proves the value on real data before you invest in a full rollout.
Have a project in mind?
Book a free 30-minute intro call. We'll talk through what you need and whether I'm the right fit, with no obligation.