Event
LA Hacks
Team
Akash Ingle, Manas Dalvi, Atharva Jayappa, Advait Atul Naik
Atlas is an on-premise multi-agent AI system built at LA Hacks 2026 that streamlines clinical workflows in emergency departments. Built by Akash Ingle, Manas Dalvi, Atharva Jayappa and Advait Atul Naik, Atlas tackles one of healthcare's most persistent problems: doctors spending nearly half their shift on documentation while patients wait on coordination gaps.
What it solves
Emergency departments are not chaotic because care is impossible. They are chaotic because information does not reach the right person at the right time. Most AI systems in healthcare are opaque black boxes that clinicians cannot trust or verify. And sensitive patient data should never leave the hospital. Atlas was built around all three of these constraints: transparent reasoning, real-time coordination, and full on-premise privacy.
What it does
Atlas coordinates five specialized agents: Scribe writes clinical notes from encounter audio, Quartermaster manages coordination and predicts delays, Advocate prepares prior authorizations during the visit, Conductor generates handoffs and discharge plans, and Reality Check audits all outputs for errors. Every claim is backed by evidence and a full reasoning chain that doctors can inspect and correct. If a premise is wrong, the system recomputes instantly. A companion Patient app lets patients review their visit summaries and answer insurance questions using only their own documents, entirely on-device.
How it works
Atlas runs locally on a single machine with a central orchestrator routing events between agents through shared memory. Every agent output must include evidence and a reasoning chain or it is rejected and rerun. Speech-to-text via Whisper powers live documentation. The mobile patient companion app is built in Kotlin with fully on-device models deployed through ZETIC, using lightweight cosine similarity for retrieval. A strict architectural boundary ensures no patient data ever crosses into any cloud system.