Hackathon · Datathon · 2026Live
An enterprise intelligence platform that pulls Jira, Git and developer activity into one knowledge graph, with AI agents on top that help managers decide who should work on what.
- 2nd
- of 31 teams at the datathon
- 40%
- less decision latency
- 4
- people on the team I led
01The problem
Engineering managers make calls every day about who should pick up what and which parts of a project are at risk. The information they need is spread across Jira tickets, Git history and people's heads, so those calls end up slow or based on gut feel.
02What we built
SynthAI pulls Jira issues, Git commits and developer activity into one knowledge graph in Neo4j, so you can ask questions across all of it at once.
On top of the graph, a multi-agent reasoning setup helps with resource allocation and software risk assessment, like spotting who already has context on a piece of code or which tickets look likely to slip.
03Decisions and trade-offs
04How it went
It cut decision latency by about 40%, and we placed 2nd out of 31 teams at the datathon.