Backend developer with 4+ years shipping production systems for EMS and healthcare — from offline AI running inside ambulances to clinical data pipelines, dispatch telemetry watchers, and Snowflake analytics platforms.
I'm a backend software developer focused on production APIs, edge AI, and healthcare data systems. At Atian.ai I architect end-to-end systems in Java/Spring Boot and Python/FastAPI — NEMSIS 3.5.1 SOAP pipelines, AI-assisted patient care report generation, offline AI deployed on ARM64 devices inside ambulances, Snowflake-backed operations dashboards, and real-time Elasticsearch infrastructure.
Before this I spent two and a half years at Tata Consultancy Services doing automation and infrastructure reliability work — the foundation for how I build now: self-healing services, offline-first design, and test suites that don't need live dependencies. Currently pursuing an M.S. in Computer Science at Jessup University.
The tools behind the systems — backend first, with a deep edge-AI and healthcare-standards streak.
Production APIs and microservices — REST, SOAP, async workflows, schema validation, JWT auth, and XML/JAXB pipelines that survive vendor quirks.
AI that runs where the cloud can't reach — quantized models on ARM64 routers, cross-compiled containers.
The unglamorous standards that make medical data move.
Production-scale automation against systems that were never meant to be automated — undocumented protocols, legacy UIs, safety-gated write paths.
Live production systems, safety-critical automation, and one meditation on civilization. Click any tile for the full case study.
Whisper.cpp + MedGemma 4B running on edge hardware inside an ambulance — real-time clinical documentation with zero internet.

Multi-tenant Snowflake platform for live ambulance districts — six tabs of real-time operations, clinical, and revenue KPIs.
Reverse-engineered a proprietary dispatch feed; self-healing watcher streaming ambulance status changes 24/7.
Safety-gated browser agent pre-filling QA-validated patient charts — every clinical decision stays human.
A meditation on civilization, disguised as a web app. Live in production.
Ambulance crews document patient encounters by hand, after the fact — and field connectivity is too unreliable (and patient audio too sensitive) to depend on the cloud. This prototype runs the entire pipeline on edge hardware inside the vehicle: an Ericsson Cradlepoint R1900 cellular router transcribes crew–patient audio with Whisper.cpp while an edge PC runs MedGemma 4B to generate clinical documentation in real time. No internet, no cloud — patient audio never leaves the vehicle.




Multi-tenant EMS operations platform for live ambulance districts. Six tabs of real-time metrics — scroll the screenshot to see the full build.
Reverse-engineered a proprietary dispatch vendor's feed and built a self-healing, always-on watcher that streams ambulance status changes into a searchable EMS data platform.
/live/health endpointA safety-gated browser agent that turns AI-extracted patient-care data into a pre-filled, QA-validated ePCR chart — cutting EMT documentation time while keeping every clinical decision human.
A meditation on civilization, disguised as a web app. Drop in a photo of any object — a pencil, a sandwich, a digital camera — and an AI vision-and-reasoning pipeline writes a monograph on what it would actually take to make it from scratch, alone: the mining, forestry, chemistry, and centuries of accumulated human knowledge behind the simplest things. Finished analyses are collected in a public "Library of Uncovered Objects," and the whole application can also run fully offline with local open-source models — adhering to the same conditions as the question itself.
Java/Spring Boot pipeline turning EMS audio into XSD-compliant SOAP submissions — GPT extraction, ICD-10/RxNorm/SNOMED mapping, on Kubernetes.
atian.ai →Python NLP pipeline scoring EMS documentation consistency — Sentence Transformers, hallucination detection, automated QA artifacts.
atian.ai →Caesar, monoalphabetic, and homophonic cipher implementations in Java.
github →Java game built on StdDraw with built-in play analytics.
github →Experiments, investigations, and what I learn while testing new technology.

A random autoplay session led me to test how convincingly AI can imitate commercial music, and what listeners are told about it.
read the investigationOpen to backend, edge AI, and platform roles — or just a good conversation about systems that have to keep working offline.