AI agents and automation
AI Systems, Agents, and Workflow Automation
Hand the repetitive work to an agent that knows your tools.
I build AI systems and automations that connect to the tools you already use and take repetitive work off your team.
Sound familiar?
The chatbot talks, and nothing happens.
It answers questions, but it can't update the CRM, route the ticket or send the invoice.
Your team copies data between tools all day.
Sales, support and finance repeat the same manual steps every day.
The prototype only works in the demo.
Real data breaks it, nobody can see why it failed, and you are not sure where your data goes.
What I build
AI features inside your product
Predictive dashboards, LLM-powered content tools and personalised experiences, built into your SaaS.
- Next.js
- Python
- FastAPI
- LangChain
- Pinecone
- GPT-4
Agents that act on live data
Support bots that run around the clock, internal task managers and lead-generation agents, sharing memory and retrying when a step fails.
- LangGraph
- CrewAI
- AutoGen
- OpenAgents
Automations across your tools
CRM workflows, ticket routing, invoices and data entry, running in the background while your team does other work.
- n8n
- Zapier
- Make.com
- Temporal.io
- Slack
- Notion
- Google Workspace
Answers from your own documents
Retrieval over your files, with logs that show what the system did and why.
- Weaviate
- Redis vector search
- OpenAI
- Vertex AI
- AWS Bedrock
Results from past work
- Top 10%on Upwork
- 100%job success score
How we would work together
We talk
A complimentary 30-minute call about the task, the systems involved and what a good result looks like.
I check that AI is the right tool
Some problems need a language model, some need a simple rule or a script. I tell you which before anything is built.
I build a narrow version first
One workflow end to end against your real tools, so you can judge it on real work.
We widen it and hand over
Once the first workflow holds up, we extend it, with notes on how it runs.
What you receive
- A working agent, automation or AI feature connected to your tools
- Orchestration with retries, shared memory and clear failure handling where it is needed
- Retrieval over your own documents when the task calls for it
- Logs that show what the system did and why
- Source code, prompts and configuration in your repository
- Documentation for running and changing it
Questions people ask about AI work
Which models and platforms do you use?
OpenAI models, Google Vertex AI and AWS Bedrock for inference, and fine-tuned open models when cost matters more than raw capability. The choice follows the task.
Can an agent work inside our existing tools?
Yes. I connect agents to tools such as Slack, Notion and Google Workspace, and to internal APIs.
What if our data is sensitive?
Tell me on the first call. Where the data goes and which model sees it is a design decision, and it is easier to get right at the start.
How do we get started?
Book the 30-minute call. You leave with clear next steps for your product and team.
