I turn business requirements into working AI: RAG assistants grounded in your documents, chatbots, agents, and integrations with your internal systems through REST APIs and MCP. Built in Python and JavaScript, from proof of concept to production.
Platforms I build on
Each one starts with a workflow discovery session and ends with something your team runs every day, connected to the tools and data you already have.
# grounded answersdocs → embed → FAISSquery → retrieve → LLM→ answer + sources
Assistants that answer from your company's own knowledge: meeting transcripts, uploaded files and Google Drive documents across PDF, DOCX, XLSX and PPTX.
# multi-turn generationuser: "draft the test cases"bot → context + history→ document / unit tests
Chatbots that draft management documents and unit test cases through guided multi-turn conversations, built into the stack your team already uses.
# agent with toolsLLM ⇄ MCP server⇄ REST API⇄ internal system
Agents that can act, not just answer: tool use, agentic workflows and Model Context Protocol servers that give AI safe access to your systems.
# secure syncOAuth 2.0 → Google Drive API→ parse → re-index# runs automatically
Connect Claude, Gemini, OpenAI and Groq to internal tools and data through REST APIs, with secure authentication and automated document synchronisation.
# prompt systemsystem prompt + template→ LLM → evaluate→ ship ↺ iterate
Production prompt systems and reusable templates that automate client business tasks, with testing, evaluation and quality checks built into every workflow.
# AI-assisted dev$ claude> plan → code → test> MCP tools • prompt patterns
Coaching engineering teams on AI-assisted development: advanced Claude Code use, prompt patterns, agentic workflows and MCP server integrations.
Full write-ups with architecture and learnings live in the AI Lab. Here is the short version.
A web chat assistant that syncs meeting transcripts from a shared Google Drive folder, answers questions about them with hybrid-search RAG, and drafts delivery documents (Executive Summary, HLRD, User Stories, Timeline Matrix) through multi-stage LLM pipelines. It is a live project, deployed through a GitHub Actions workflow.
A multi-agent orchestration system that works as a small brand's virtual executive team: one conversational orchestrator routes every request to six specialist agents, who produce campaign strategy, launch and event plans, roadmaps and print-ready apparel artwork. A full executive-team run takes about 60–90 seconds and $0.10–0.30.
I sit between technology, product and people: I can scope the problem with a department head and write the integration with the engineers.
You know the workflow pain. I find where AI actually helps.
You own the systems. I help AI fit into them cleanly.
Short cycles, a working proof of concept early, and documentation that lets your team own the result.
Workflow discovery with stakeholders to find the pain points worth solving.
Check whether AI fits, pick the platform and plan the integration.
Build a working proof of concept on real data in a short cycle.
Connect it to internal systems through APIs, OAuth and MCP.
Setup guides, usage guidelines and coaching so the team owns it.
Feedback from the prototype loops back into the assessment before anything is scaled.
Chains, agents and multi-agent workflows, connected to real systems through the Model Context Protocol. The same frameworks I teach in institute cohorts.
Enablement, solutions and training under one roof.
Chatbots, RAG pipelines, agents, automation and REST API integrations that connect AI to the systems your team already uses.
Solutions profile →Workflow discovery, reusable prompt libraries and AI skills, and documented usage guidelines so adoption stays secure and consistent.
Enablement profile →Hands-on workshops for mixed audiences, so business users and technical staff both leave able to use AI on their own work.
See workshops →A software engineering career since 2013: enterprise apps, microservices and now AI systems.
Discovery with stakeholders first, so what gets built solves the problem the team actually has.
OAuth 2.0, VPN-secured tooling, data handling standards, and systems built for healthcare and pharma.
Docs for end users and engineers, plus coaching, so the solution doesn't depend on me.
Tell me what your team is trying to do. I'll show you where AI fits and what a first proof of concept would look like.
Claude and the Anthropic API, Google Gemini, OpenAI and Groq, plus AI coding environments such as Claude Code, Cursor and Google Antigravity. I build on whichever platform your organisation has approved, and I can integrate any other model or tool you already use in your environment.
Yes. Integrations run through REST APIs, the Google Drive API, OAuth 2.0 and MCP servers, so AI works with your data where it already lives.
Secure authentication, data handling and acceptable-use practices are part of the design, including builds for VPN-secured environments.
Setup guides, usage guidelines and technical handover documentation for both end users and engineers.
Yes. I coach teams on Claude Code and AI-assisted development. See the training profile for workshops.
Every engagement is scoped to your team's needs, so pricing depends on what we build together. Tell me about your project and I'll share a tailored proposal.