Frontend Engineer, Chanakya

Sarvam is building what it calls the bedrock of sovereign AI for India, developing a full stack platform across research, models, infrastructure and applications, backed by Lightspeed, Peak XV and Khosla Ventures and working with brands including Tata Capital, SBI Life, CRED, IDFC and LIC. This role is on Chanakya, the team serving strategic sector and complex enterprise accounts, and the posting is blunt that these are not form and table frontend roles. You would build the interfaces sitting directly on top of Sarvam's AI stack: geospatial overlays, simulation interfaces, natural language access layers, ontology viewers, data ingestion dashboards and complex enterprise workflows. The defining constraint is the environment. These interfaces have to work offline first, on low bandwidth, on limited hardware and inside hardened client machines, not just on a developer's laptop over broadband. You would own the frontend layer of what Sarvam calls atoms, which are replicable deployable product components, from architecture through to deployment at client sites, and take part in field testing.

What the posting requires

- 3 to 6 years in frontend engineering with shipped, production grade products
- Deep React proficiency, TypeScript, modern CSS such as Tailwind, and strong component architecture instincts
- State management with Zustand, Redux Toolkit or equivalent, plus solid async data patterns
- Data visualisation experience with at least one of D3.js, Mapbox or Leaflet, Deck.gl for geospatial, or Recharts, used in production
- REST and WebSocket integration, with enough backend awareness to debug across the boundary
- Performance engineering: bundle optimisation, lazy loading and rendering performance in low resource environments

Counts as a bonus, not a requirement

- WebGL or Three.js for 3D visualisation
- Electron or desktop application experience
- Offline first or progressive web app architecture
- Prior work on dashboards, command and control interfaces, or complex enterprise tooling

What you would actually be doing

- Building production grade interfaces for strategic sector and enterprise AI use cases, including geospatial overlays, simulation interfaces, ontology viewers and data ingestion dashboards
- Owning the frontend layer of atoms end to end, from component architecture through deployment into client embedded environments
- Building for genuinely constrained conditions: offline first, low bandwidth and limited hardware
- Contributing to shared frontend infrastructure, component libraries, the design system and state management patterns
- Joining field testing and client feedback loops alongside the deployment engineers

Worth knowing before you self reject

Sarvam adds a note to this posting saying it is looking for people who can own the outcomes described, not people who match every line of the specification. If the problem interests you and you believe you can do the work, the company is explicitly inviting you to apply.

A good fit if

You are a frontend engineer who wants to work close to real AI systems rather than on marketing pages, and you find hard constraints interesting rather than irritating.

Think twice if

You are used to shipping to modern browsers on fast connections and would rather not think about offline behaviour or locked down client machines. That constraint runs through the entire role.