anashitanema.dev

Senior Software Engineer · Nagarro

Ashita
Nema.

I build Flutter apps for iOS, Android and the web, with a focus on clean architecture, design systems and server-driven UI. Lately I've been connecting mobile apps to multi-agent AI on Google Cloud.

Get in touch GitHub LinkedIn
  • 4+ years at Nagarro
  • Top 100 · Patchamomma 2026
  • Speaker · Nagarro FLO 2026
  • Madhya Pradesh, India
lib/about.dart
final ashita = Engineer(
  role: 'Senior Software Engineer',
  company: 'Nagarro',
  since: 2022,
  stack: [Flutter, Dart, Bloc],
  loves: [
    'clean architecture',
    'design systems',
    'server-driven UI',
  ],
  building: 'llm_markdown_renderer',
);

// Hot reload is still my favourite feature.

featured project

Top 100 of 3,500 engineers · Google Cloud Patchamomma 2026

CivicPulse

Report. Predict. Verify.

A Flutter app backed by five AI agents on Google Cloud. A citizen photographs a pothole or overflowing bin. The system classifies it, merges duplicates, routes it to the right department, checks that the fix really happened, and predicts which wards will need attention next.

Role
Flutter engineer
Built
September 2026
Platforms
iOS, Android
Cloud
Google Cloud
CivicPulse home screen with Report an Issue and View Reports buttons
Home
Report screen with a pothole photo and auto-captured GPS coordinates
Photo + GPS, no forms
Report submitted: classified as a critical pothole with 98% AI confidence
Classified by Gemini
Reports feed showing pothole tickets for Ward 8 with severity and status chips
Live ward feed

The CivicPulse loop

Most civic apps stop at report, route, mark resolved. This one keeps going.

  1. ReportTake a photo and tap submit. GPS is captured automatically, with no category to choose.
  2. ClassifyGemini 1.5 Flash returns the issue type, severity, a confidence score and a short description.
  3. Deduplicate and routeOpen reports of the same type within 200 m are merged. A routing agent assigns the municipal department.
  4. VerifyWhen a ticket is marked resolved, the citizen adds a follow-up photo. Gemini compares before and after, and flags fake fixes as false_closure.
  5. PredictA daily BigQuery ML job forecasts which wards are likely to break next and publishes them to a public Looker Studio dashboard.

Architecture

Event-driven and serverless. Every verified fix feeds back into the forecasting model.

Flutter app→ Firestore + Storage→ Pub/Sub→ Cloud Run · FastAPI→ Google ADK
Vision classifier Duplicate detection Routing Verification Insight (BigQuery ML)
Where will the next problem happen?A forecasting model trained on report density, issue type and season flags high-risk wards before anyone reports them.
Did the fix actually work?Semantic before/after comparison catches false closures instead of trusting a status change.

Demo: reporting a pothole end to end, from the app to Firestore and the Looker Studio dashboard3:43

open source & speaking

Things I share

llm_markdown_renderer

Published on pub.dev · verified publisher ashitanema.dev

A Flutter package that renders Markdown streaming in from an LLM, token by token. It parses incrementally, so long answers don't re-render from scratch, and half-finished code blocks and tables stay stable while they stream.

flutter pub add llm_markdown_renderer
pub.dev ↗ All my packages ↗ GitHub ↗

Talk · Nagarro FLO 2026

Domain-specific AI

Presented a session on domain-specific AI at FLO, Nagarro's annual event, on applying AI to the language, data and workflows of a particular business domain.

experience

Nagarro

Gurugram · Jul 2022 – now

  • Senior Software EngineerJun 2025 – now
  • Software EngineerJan 2024 – 2025
  • Associate Software EngineerJul 2022 – Jan 2024
  • Build an enterprise Flutter app on clean architecture with BLoC, structured so features can be added without touching unrelated layers.
  • Work on server-driven UI and the design system behind it: shared tokens, theming and a library of reusable widgets.
  • Built proofs of concept for new client work, including several on Flutter Web.
  • Ship features across fintech, enterprise workflow and consumer apps with product, design and backend teams.
R&D Intern · ETHNUSSep – Nov 2021 · Bengaluru

Research and development for AWS recorded training.

Angular Developer Intern · ITESep – Nov 2020 · Bhopal

Front-end development with Angular.

B.Tech, Computer Science2018 – 2022

Rajiv Gandhi Prodyogiki Vishwavidyalaya

skills

What I work with

Mobile

  • Flutter, Dart
  • BLoC, Riverpod, Provider
  • Clean architecture
  • Retrofit, Dio, REST
  • Native platform integration
  • iOS + Android releases

UI systems

  • Design systems
  • Server-driven UI
  • Theming and design tokens
  • Flutter Web
  • Reusable widget libraries

Cloud & AI

  • Gemini (multimodal)
  • Google ADK multi-agent
  • MCP Toolbox for Databases
  • Cloud Run, Pub/Sub, Scheduler
  • Firestore, Cloud Storage
  • BigQuery, BigQuery ML, Looker Studio

Backend & web

  • Java, Spring Boot
  • Python, FastAPI
  • Node.js
  • Angular
  • Web development

contact

Get in touch

Email is the fastest way to reach me.

LinkedIn GitHub