AI Research & Fact-Checking System
A research agent that decomposes a claim, gathers evidence from multiple sources, cross-checks agreement between them, and refuses to answer when the evidence is thin.
Python · LangGraph · Claude API · PostgreSQL
AI · Data · Automation
Learn Data, AI and Automation by building real projects — from business analytics and AI agents to modern data platforms and algorithmic research.
Featured build
A research agent that decomposes a claim, gathers evidence from multiple sources, cross-checks agreement between them, and refuses to answer when the evidence is thin.
Python · LangGraph · Claude API · PostgreSQL
Real systems, documented end to end — architecture, code, trade-offs.
A research agent that decomposes a claim, gathers evidence from multiple sources, cross-checks agreement between them, and refuses to answer when the evidence is thin.
Most RAG systems are tuned by vibes. This is a reproducible harness that measures retrieval and generation separately, so you can see which half is failing.
An inventory dashboard built around the four decisions a stock controller makes each week, with a semantic model that stays fast on 40 million rows.
Most RAG evaluations measure the wrong thing, on the wrong data, with a judge that rewards the wrong behaviour. Here is what to measure instead.
Backtests fail in a consistent direction — optimistic. Here are the five structural reasons, with a detection method for each. Educational research only.
A staged plan for moving from report development into platform engineering on Microsoft Fabric — including an honest read on which existing skills carry over.
A practical decision guide based on workload shape rather than benchmark scores — because the fastest tool depends on what you are actually doing.
Pick a destination, not a playlist.
From prototype to a language-model system that survives production
7 steps · ~12 weeks
Turn ambiguity into decisions, with AI doing the mechanical work
4 steps · ~8 weeks
SQL, analysis and a portfolio that gets interviews
5 steps · ~10 weeks
Pipelines, modelling and orchestration — starting on your laptop
5 steps · ~14 weeks
From report developer to platform engineer
6 steps · ~14 weeks
Automations that still work in six months
4 steps · ~8 weeks
Understand it. Code it. Backtest it. Honestly.
6 steps · ~12 weeks
Model time saved, error reduction, build cost and ongoing maintenance to get a payback period you can defend.
Open toolModel input and output tokens, request volume and caching to project monthly cost across model tiers.
Open toolPaste a return series and get the risk-adjusted metrics that matter, with the assumptions stated. Educational only.
Open toolWork out how many chunks a corpus produces, what retrieval depth fits your context window, and what indexing will cost.
Open toolShort problems that build real judgement.
Produce a running total of amount per customer that resets at the start of each calendar month, ordered by transaction date.
Try this question20 questions across 7 tracks.
Interview LabHypotheses, methods, code and results — reproducible.
Testing the classic RSI(14) mean-reversion rule across 20 years and 500 instruments, with costs included and multiple-testing accounted for.
A reproducible benchmark across six workload shapes, reporting memory as well as time — because memory is what actually stops the job.
Fixed, recursive, semantic, structural, sentence-window and parent-document chunking measured on retrieval recall across three corpora.
Retrieval that holds up, agents that fail safely, evaluation that catches regressions, and MCP for tool integration.
Process design, idempotency, error handling and observability — the engineering discipline that separates a workflow demo from a production system.
Move from building reports to building the platform reports sit on: lakehouse architecture, medallion layering, Direct Lake and deployment pipelines.
Community
Build alongside other engineers, analysts and founders working with data and AI. Weekly builds, code reviews and a growing resource library.
Weekly builds · code reviews · resource library · live sessions
One email a week. Source code and files included. No fluff, no recycled LinkedIn posts.
Business Automation Audit · Power BI Dashboard Development · Microsoft Fabric Implementation · AI Workflow Development