Skip to content
Services

Development services, end to end

AI feature work and the web or mobile software it sits in. Public notes on the blog show the patterns I use (retries, streaming, evaluation sets, prompt-injection defenses, pgvector). I do not invent client results here.

01

LLM feature integration

Adding LLM features to web apps: model APIs, streamed responses, retries, JSON validation and cost control. The patterns are the ones I have written up in public notes on the blog.

  • Streaming responses into a React UI (server-sent events)
  • Retries and error handling around LLM API calls
  • Basic cost and token-budget checks
  • JSON validation of model output
02

AI agents & automation

Workflow bots and tool-using agents with guardrails, including checks that stop a model from calling a tool it should not. Python automation sits alongside this, from data scripts to app backends.

  • Tool-gate checks for model-initiated actions
  • Workflow bots around existing APIs
  • Python automation and data-processing scripts
03

RAG / search over company data

Search and retrieval over your documents or records, using Postgres/pgvector or a similar store, with a small evaluation set so you can tell whether answers actually improved. Related (non-LLM) search work includes a healthcare provider search tool built with Django APIs and Mapbox.

  • Retrieval over a document or records collection
  • Postgres/pgvector or a comparable store
  • A starter evaluation set so changes are measured
04

AI web & mobile MVPs

End-to-end web and mobile application development, with AI features where they fit.

  • React front-end application development
  • Node.js and Python backend API development
  • Mobile app MVPs (iOS and Android foundation since 2015)
  • AI-assisted features wired into the product where they help
05

Local AI / hardware setup advice

Advice on running models locally and on choosing a cloud GPU or a local machine. This is advisory; I do not sell hardware. It ties to the buying guides I am building on the blog.

  • Help choosing between a local setup and a GPU cloud
  • A practical local-model layout for the machine you already have
  • Notes you can reuse when a tool or GPU comes up for purchase

Questions

What is included in LLM feature integration?

I add LLM features to web apps using model APIs, streamed responses into a React UI with server-sent events, retries and error handling around those API calls, JSON validation of model output, and basic cost and token-budget checks. These are the patterns I have written up in public notes on the blog.

How do you build AI agents and automation?

I build workflow bots and tool-using agents with guardrails, including checks that stop a model from calling a tool it should not. That work covers tool-gate checks for model-initiated actions, workflow bots around existing APIs, and Python automation and data-processing scripts alongside app backends.

How do you add search over company data?

I build search and retrieval over your documents or records, using Postgres/pgvector or a similar store, and a small starter evaluation set so you can tell whether answers actually improved. Related, non-LLM search work includes a healthcare provider search tool built with Django APIs and Mapbox.

What does an AI web or mobile MVP include?

It is end-to-end web and mobile application development, with AI features wired in where they help. That typically covers a React front end, Node.js and Python backend API development, and mobile app MVPs, on an iOS and Android foundation I have had since 2015.

Do you sell hardware, or only advise on local AI setups?

This service is advisory; I do not sell hardware. I help with running models locally, choosing between a local setup and a GPU cloud, a practical local-model layout for the machine you already have, and notes you can reuse when a tool or GPU comes up for purchase. It ties to buying guides I am building on the blog.

Projects

Selected projects

Automated market makers, an on-chain PvP game, arbitrage bots, a DEX fork and a healthcare search tool. Client names are withheld under NDA.

Full-Stack Web & GIS NDA

Healthcare Provider Search Tool

Built for a US hospital

Medication-provider search tool for a US hospital using a React and ReduxForm front end, Mapbox maps and Django REST APIs.

  • Provider search for a US hospital
  • React and ReduxForm front end
  • Mapbox integration
  • Django REST APIs
ReactReduxFormMapboxDjangoDjango REST Framework (DRF)Python
Technologies

Tech stack

Smart contracts, Web3 integration, backend and frontend — the tools behind every engagement.

Backend Development

Node.js

Backend services, API integrations, and server-side development.

Frontend Development

React

Web application front ends and application interfaces.

AI, data & automation

Python

Python scripting, data analysis, automation, and API backends.

Also used in projects
  • Django / DRF
    Used in a healthcare provider search tool built for a US hospital.
  • ReduxForm
    Form state handling in the healthcare provider search tool.
  • Mapbox
    Map integration in the healthcare provider search tool.
  • LLM APIs, SSE, pgvector, evals
    Topics of public notes and in-browser demos on this site. Confirm with me which of these I have used in a shipped client project.
> ready_to_build()

Have an AI feature in mind?

Tell me what you want to add or ship. I will reply with scope, approach and a quote.