SDE 2

SDE - 2
Type: Full time
Location: Electronics City, Bangalore
Experience: 3-5 years
About Us: We build cost engineering software for manufacturers. Our customers use it to work out what a part should cost before they negotiate a price for it.
We are looking for one person to own engineering for both products end to end: the code, the team, the release process, and the security of a platform that holds commercially sensitive data for multiple customers at once.
What you would be taking on
Two products, four repositories, two backend services and two frontend deployments of a shared codebase. Roughly 1,160 HTTP endpoints across the backends. A multi tenant architecture where every customer gets their own database and their own identity realm, with a shared reference database holding costing knowledge that all customers draw from.
Python 3.12, FastAPI 0.115, Tortoise ORM 0.25, aerich, PostgreSQL, Redis, Celery 5.5, Temporal, Keycloak,
Next.js 15.5, React 19, TypeScript 5, MUI 7, AWS S3
The platform has real engineering depth. The costing engine is configuration driven, so cost models are database rows rather than code and a new manufacturing process can be modelled without a release. There is a durable workflow engine running a document extraction pipeline. There is a sensitivity and scenario engine for what if analysis. Multi tenancy is done properly, with isolation at the database, storage and identity layers rather than row level filters.
What you would own1. Delivery
Own the release train across four repositories and a branch per environment model.
Plan and sequence work against the product roadmap, and be accountable for it landing.
Keep the two frontend deployments of one shared codebase from diverging.
2. The team
Run a team of engineers across backend, frontend and QA. Set expectations, give feedback, handle performance honestly.
Hire. Write the specs, screen, interview, and sell the role. You will be the main technical voice in the loop.
Onboard people into a large codebase and make them productive quickly.
3. Engineering practice and SDLC
This is where the biggest gap is today, and where you would have the most immediate impact.
There are around 184 backend test files and 34 frontend test files already written. None of them run automatically. There is no continuous integration in any of the four repositories.
There are no container definitions and no infrastructure as code. Deployment is operated outside the repositories.
Database migrations are not consistently under version control, so environment drift is invisible until it causes a problem.
You would define the code review standard, the branching model, the definition of done, and the quality gates. Then make them stick.
4. Security
Own the security posture of a multi tenant platform holding commercially confidential customer data.
Own the remediation backlog from our internal security testing, and the process that keeps it from growing back.
Secrets handling, dependency hygiene, access control, tenancy isolation, and the controls that let a customer audit us.
Build security into the pipeline rather than treating it as a periodic exercise.
5. Architecture
Make and document the calls on data modelling, service boundaries, asynchronous processing and caching.
Keep the platform scaling as we add customers, with a close eye on anything that gets slower as tenant count grows.
Decide what gets paid down and what gets lived with. You will have to say no to some things.
How we would measure the first yearFirst 30 days
You can run the whole stack locally and have shipped something small to production.
You have read enough of the codebase to tell us what we have got wrong.
You have met the team one to one and formed a view.
First 90 days
Continuous integration running on every pull request in all four repositories, with the existing tests actually executing.
Container definitions committed, so a new engineer or a new environment starts from the repository.
Migrations under version control with a documented path to apply them.
The security backlog triaged and sequenced, with the top items closed.
A hiring plan agreed, with a pipeline running.
First year
Releases are routine rather than events, and we can say what is in production with confidence.
Test coverage is meaningful on the costing engine, which is the part of the product customers pay for.
The team has grown and new joiners reach productivity in weeks.
We can answer a customer security questionnaire without a scramble.

Work arrangement
No