What I build
Quick prototype to validate the idea, or a production system your team can run without me. You own everything.
AI Agents & Automation
Someone on your team is making the same judgment calls 50 times a day — triaging tickets, qualifying leads, processing invoices, routing approvals. They've been asking for help since last quarter.
Agents that handle the repetitive decisions — with monitoring so you know they're actually getting it right.
- → Multi-agent orchestration (LangGraph)
- → Internal ops — IT, HR, finance workflows
- → Customer-facing — onboarding, support, product
- → Tool use and system integrations
- → Human-in-the-loop escalation
- → Evaluation and observability from day one
RAG & Knowledge Systems
Your analysts spend days compiling what should take minutes. The answers are buried across reports, filings, and docs nobody reads twice.
- → Document ingestion and chunking pipelines
- → Hybrid retrieval (vector + keyword)
- → Structured extraction — compliance docs, reports, regulatory filings
- → Source attribution on every answer
- → Accuracy monitoring and evaluation
AI Automation & Data Enrichment
Half your CRM is incomplete. You're managing credits, failed lookups, and quarterly data fire drills — or doing it by hand, one record at a time.
- → LLM enrichment with web search grounding
- → CRM integrations (Salesforce, HubSpot)
- → ICP scoring with reasoning, not just numbers
- → Scheduled runs via GitHub Actions
- → Unlimited records — no credit system
AI Strategy & Use Case Identification
Everyone in the company has an AI idea. None of them have been evaluated. Your vendor is selling you their tool, not the right tool.
- → Use case discovery and prioritisation
- → Build vs buy analysis
- → Vendor and tool evaluation
- → Architecture review
- → Roadmap with realistic timelines