skip to content
akansha:roy

about

a bit about me

i got into programming by building things, and somehow kept going down rabbit holes i didn't know existed. these days, i'm mostly into backend and systems, with ai being the next thing i'm figuring out.

a little about me

i got into programming in 2025, starting with node.js and full-stack projects. even back then, i was more interested in building something of my own than following the usual project tutorials. my first few projects were pretty different from each other, from a skincare platform built around home remedies to dashboards for managing freelance work and a small airbnb-style platform. looking back, i was basically a kid with a new toolbox, trying to build whatever came to mind lol.

working on these projects eventually got me closer to backend development. during my internship, i started spending more time on the backend side of things and got interested in what happens behind an application rather than just what it does on the surface.

how i learn

i've never been great at learning things just by reading about them. i usually need to build something, run into a problem i don't understand, and then go figure out what's happening.

that's how i got into systems. i was learning gRPC when i decided to try building a small rpc framework of my own. later, i built a protobuf compiler to support it. i also tried implementing a kafka-inspired log, which introduced me to concepts i hadn't really explored before. one project kept leading to another, and i often ended up going much deeper into a topic than i'd originally planned.

i think that's what i enjoy most about building from scratch. you start with a rough idea of how something works, and somewhere along the way, you discover just how much you didn't know. i like that process, even when it gets frustrating.

where my interests are going

systems are something i keep coming back to, almost regardless of what i'm working on. i tend to get curious about the layer underneath, how something is implemented, why it behaves the way it does, and what happens when things don't go as expected.

lately, i've been exploring the space between backend engineering and ai. i want to understand llms beyond just calling an api, build agents of my own, and figure out how these applications fit into the backend systems i already enjoy working on.

i'm still figuring out which parts of ai i'll enjoy the most, and i think that's part of the fun. i want to keep building, keep asking questions, and see where that takes me.

outside of it

outside of code, i like drawing, cooking, art and making things in general. i read manga and manhwa fairly often. i pick up technical books sometimes too, but i am not a regular book reader.

01curiosity

currently curious about

02experience

where i have worked

Aug 2026 — Sep 2026

DiligenceAISoftware Engineer (Contract)

TypeScript, PostgreSQL, AWS, Hono, Next.js, GitHub Actions, Docker

  • built external API integrations for authentication and transaction sync, with credential lifecycle handling, durable checkpoints, bounded retries and recovery from provider failures.
  • implemented PostgreSQL models, migrations and regression tests for sync state; improved background jobs with worker leases, queue deduplication and structured error reporting.
  • added GitHub Actions and Docker checks for tests, types and builds, and worked on AWS staging and deployment workflows.

Jul 2025 — Oct 2025

Repro India LimitedSoftware Developer Intern

Java, Spring Boot, Thymeleaf, SQL, BIRT, AWS S3

  • built a scheduled validation job that checked incoming orders and corrected combo pricing and page-count logic that had been wrong in a few edge cases.
  • built a bulk invoice download endpoint backed by S3, which turned a two-hour manual export into a single request.
  • reworked the SFTP ingest workflow so failed uploads surface properly instead of leaving files behind silently.
03find me

elsewhere on the internet

email is the best way to reach me. i am based in India.