Mobile apps
Native Kotlin & Jetpack Compose apps: fast, offline-first, clean architecture and store-ready.
- Kotlin & Compose
- Offline-first
- Play Store release
Open to work · Pune, India
I'm Eshwar. I build Android apps, REST APIs and cloud backends end to end, from the first commit to production, and add AI where it helps.

2+ yrs
building products
Kotlin · Ktor · GCP
React · Gemini AI
2+ yrs
Professional experience
4
Platforms: mobile, backend, web, cloud
3
Open-source projects on GitHub
8.5
M.Sc. Computer Science CGPA
What I do
One engineer across the stack: the app people use, the API behind it, the cloud it runs on, and the AI that makes it smarter.
Native Kotlin & Jetpack Compose apps: fast, offline-first, clean architecture and store-ready.
Ktor REST APIs on PostgreSQL with auth, caching and clean docs, built to be easy for mobile and web to consume.
Deploy and run services on GCP: containers on Cloud Run, managed databases, storage, and CI/CD that ships on every push.
Add Gemini-powered features to apps: smart assistants, summaries and chat grounded in your own data.
Experience
Shipping production software for real users, from internship to full-time engineer.
Dec 2024 to Present
VKS Infotech Pvt. Ltd. (Anireysoft), Pune
Sep 2024 to Nov 2024
VKS Infotech Pvt. Ltd. (Anireysoft), Pune
Education
M.Sc. Computer Science
Advanced algorithms, software engineering, mobile development
2023 to 2025 · CGPA 8.5
B.Sc. Computer Science
Padmashri Vikhe Patil College, Pravaranagar
2020 to 2023 · CGPA 8.2
Projects
Open-source projects you can read, run and fork on GitHub.
suspend fun complete(purpose: String, prompt: String)
: Completion {
for (step in chain(purpose)) {
val p = providers[step.provider] ?: continue
runCatching { return p.complete(prompt, step.model) }
}
error("all providers failed")
}AI Backend · Kotlin
Ktor service that puts one interface in front of several LLM vendors (Gemini, Claude). Each purpose gets its own model and failover chain, with retries on transient errors and per-call token and cost telemetry.
async def ask(self, question: str) -> Answer:
vec = self._embedder.embed(question)
hits = self._store.search(vec, k=5)
text = await self._generate(question, hits)
return Answer(text, hits, grounded=bool(hits))RAG · Python
Ask questions about your own PDFs and get answers with citations to the document and page. Boundary-aware chunking with overlap, Gemini embeddings, SQLite vector search and grounded answers served by FastAPI.
<Reveal>
<h2>Turning complexity
into <Em>clean software.</Em>
</h2>
</Reveal>Web · Frontend
This site: a Next.js and Tailwind CSS portfolio with scroll and mouse-driven animations, statically exported and deployed on Cloudflare with automatic deploys from GitHub.
Skills
The languages, frameworks and platforms I use every day.
Native Android apps with clean, testable architecture.
Fast, secure REST APIs and data layers.
Responsive dashboards and admin panels.
Deploy, scale and monitor in production.
Practical AI features inside real products.
Tools for building, testing and shipping.
Contact
Hiring for a role or planning a product? Send a message and I'll reply within 1 to 2 days.