SignalBrief

Why I created SignalBrief.

I wanted a project that was closer to real analyst and AI engineering work than a standard modelling demo. SignalBrief shows how I think: collect evidence, structure uncertainty, expose risk, and turn messy inputs into a decision-ready brief.

Messy public informationNews, filings, blogs, analyst notes, customer comments, and raw search results.
→
Agentic research workflowPlanner, searcher, RAG reader, extractor, analyst, and writer nodes.
Unclear signalUseful facts are mixed with noise, repetition, and weak source quality.
→
Structured decision outputRisks, opportunities, metrics, sentiment, citations, and JSON export.

It matches the roles I want

Consulting, hedge fund, enterprise AI, and data teams all need people who can go from information overload to grounded analysis quickly. This project mirrors that daily workflow.

It combines my strongest skills

The build brings together Python, LangGraph, RAG, structured output, testing, UI polish, and product thinking instead of isolating one small technical trick.

It is interview-friendly

The architecture is easy to explain: each agent has a clear responsibility, every output section has a schema, and the product can run instantly without secrets.

What I wanted recruiters to see

Can build AI systemsLangGraph
Can ground output in sourcesRAG
Can think like an analystRisks
Can ship product-quality workStreamlit

Why not another time-series project?

Time-series modelling can be useful, but it is common on CVs and often looks like a notebook. SignalBrief is more differentiated because it shows orchestration, retrieval, product design, structured outputs, and decision-support thinking in one place.

That is closer to the type of work firms actually pay for: systems that make skilled people faster.

AI EngineeringAgent workflow, RAG, structured outputs, provider-ready search.
Data EngineeringClean models, repeatable pipeline, exportable JSON, tested core.
Analyst ThinkingRisks, opportunities, sentiment, metrics, and evidence quality.
Product PolishPremium UI, preview site, process page, and recruiter-friendly docs.