About Kalpla

At Kalpla, we believe learning should feel engaging, personal and within reach for every student. We are building a platform that supports students through their entire learning journey, and gives instructors the tools to share recorded courses and teach live with confidence. We are a small, ambitious team, and we improve the platform every day to get closer to that vision.

Why this role is engaging: a Research & Development role

Most developer jobs at this experience level hand you a ticket with the design already decided. This is not that job.

Our Product

Manager will bring you a feature at the level of "instructors need a way to run a quick poll during a live class".

Everything after that sentence is yours to work out: how it fits into our system, which approach is right, what it costs to run, how it behaves when things go wrong, and what we build first. You will research the options, weigh the trade-offs, recommend the best architectural decision, and then build it end to end. You will be working on something new almost every day.

What a typical feature looks like here

1. Our PM shares the high-level intent and the business reason behind it.
2. You research: how our current system works, how others have solved the problem, which libraries or services fit, and the constraints (scale, cost, security, real-time behaviour).
3. You come back with a short technical design: the approach you recommend, the options you rejected and why, and the risks.
4. We align as a team, then you build it across the database, APIs and interface, with AI as your pair.
5. You test it properly, ship it to staging, and see it through QA to production.

We value your reasoning in step 3 as much as your code in step 4. What you will work with

- Frontend: Next.js 15 (App Router), React 19, TypeScript, Tailwind CSS, TanStack Query, Zustand, react-hook-form and Zod
- Backend: NestJS, PostgreSQL with Prisma, Redis, Keycloak, and a shared API contract package
- Real-time: live classes on a mediasoup SFU server, mediasoup-client and Socket.io signalling
- Platform: AWS (ECS, S3, Lambda, CloudFront), Vercel, GitHub Actions, pnpm + Turborepo monorepo

Roles and responsibilities

- Technical research and design: turn a high-level feature brief into a defensible architectural decision. Explore options, prototype where it helps,



and write up your recommendation with its trade-offs.
- Frontend: build features in Next.js, choosing the right rendering approach (server components, SSR, client-side) with SEO and performance in mind. Turn Figma designs into responsive, accessible UI with proper loading, empty and error states.
- Backend: build REST APIs in NestJS using modules, services, dependency injection, validated DTOs and guards. Design and query data in PostgreSQL through Prisma, including relations, indexes, transactions and safe migrations. Keep the shared API contract in sync and integrate third-party services.
- Live classes: work on our real-time video system and debug real-world issues such as joining, audio and video quality, reconnects and recordings.
- Quality: write unit, integration and end-to-end tests, add a regression test for every bug fix, and take part in code review.
- Security: enforce authorization on every endpoint, validate all input on the server, never trust client-sent values (especially payment amounts), and handle student personal data and secrets responsibly.

AI-assisted development (Claude Code) We develop using Claude Code, so you will work alongside an AI coding assistant every day. We want someone who uses AI to move faster without giving up ownership, understanding or quality.

- Give the AI precise context: the goal, constraints, relevant files and what "done" looks like. Break large work into well-scoped tasks
- Plan before building, and review the approach before any code is written
- Read and understand every line the AI writes. You own the code, not the tool
- Verify, do not assume: run the app, write and run tests, check edge cases
- Spot when the AI is wrong: invented APIs, outdated patterns, or fixes that treat symptoms instead of causes
- Apply extra scrutiny to payments, authentication and database changes
- Keep secrets and personal data out of prompts, and help improve our AI setup (project context files, reusable workflows, team conventions)





Preferred candidate profile

- 2 to 4 years building production web applications with TypeScript
- Solid React experience and hands-on Next.js (App Router preferred)
- Hands-on backend experience with Node.js. NestJS preferred; Express is fine if you are keen to learn NestJS
- Strong SQL and relational database fundamentals (PostgreSQL preferred) and experience with an ORM
- Clear understanding of REST APIs, HTTP, authentication versus authorization, and asynchronous programming
- Able to turn an ambiguous, high-level requirement into a concrete technical design, and explain trade-offs in cost, complexity, performance and maintainability
- Hands-on experience with an AI coding assistant (Claude Code, Cursor, GitHub Copilot or similar) in real projects, with the judgement to review and own AI-generated code
- Clear written communication, comfort with ambiguity, and comfort with Git, pull requests and code review

Nice to have

- Claude Code specifics: CLAUDE.md project context, plan mode, subagents, skills or custom commands; familiarity with MCP
- WebRTC fundamentals, or experience with mediasoup or another SFU; Socket.io or WebSockets
- Prisma, Redis, Keycloak, OAuth2 or OIDC; payment gateway integration (Cashfree, Razorpay or similar)
- Docker, AWS, Vercel and CI/CD with GitHub Actions; EdTech experience

Who this role is probably not right for

- If you prefer receiving detailed specs and implementing them exactly, this role will feel unstructured
- If you want to specialise only in frontend or only in backend, the constant switching will be tiring
- If AI-assisted development feels like a threat rather than a tool, we are probably not aligned

Perks and benefits

- Real ownership: your name is on the architectural decision, not just the pull request
- New problems almost every day, across web, backend and real-time video
- Direct access to the Product Manager and the Head of Engineering
- A team genuinely investing in AI-assisted development
- A live product with real students and instructors

Work model: On-site in Udupi for the first 45 days, then hybrid. Hiring process: Profile review, intro conversation with the Head of Engineering, a practical task (Claude Code or any AI assistant welcome), a walkthrough where you explain your design and code, then the offer.

📌 Full Stack Developer (R&D) (Udupi)
🏢 Learncap Academy
📍 Udupi