Website:
sarvam.io
Job details:
About Sarvam
Sarvam is building the bedrock of Sovereign AI for India. The company is developing India's full-stack sovereign AI platform, building across research, models, infrastructure and applications with a singular focus on making AI genuinely work for India. Sarvam works with leading enterprises and public institutions and is backed by Lightspeed, Peak XV, and Khosla Ventures. Sarvam partners with India's leading brands, including Tata Capital, SBI Life, CRED, IDFC, and LIC.
About The Team
Sarvam's research teams build our own vision-language models for OCR and structured extraction. This team builds everything around them — the serving harness that turns a 3B or 30B in-house model into a production document intelligence platform.
The bet is specific: with the right harness — routing, decomposition, retries, verification, ensembling, layout awareness, confidence calibration — a small sovereign model should match or beat what teams today get from frontier hosted models like Gemini Flash, at a fraction of the cost and fully within India. Closing that gap is an engineering problem, and it is this team's problem.
We run against the full messiness of Indian documents at population scale: PAN and Aadhaar, bank statements, GST filings, insurance and medical reports, 60-page
contracts, legal filings and RFPs — across languages, scan quality, and layouts that were never designed to be machine-read.
Stack: Go, Python, Temporal, REST, Kubernetes, PostgreSQL, Redis, object storage, OpenTelemetry-based observability.
About The Role
Document intelligence is only trustworthy if a human can see what the model saw. You will build the interfaces where that happens: document viewers with field-level grounding, review and correction workflows, extraction schema builders, evaluation dashboards, and the developer-facing console our enterprise customers use to run pipelines at scale.
These are dense, stateful, performance-sensitive UIs — rendering 100-page PDFs with overlaid bounding boxes, streaming results as pages complete, letting a reviewer correct a field and push that correction back into the loop. The quality of this surface directly determines how much our customers trust the system.
You will be the frontend owner for the team, working closely with backend and applied AI engineers rather than against a finished spec.
What You'll Do
- Build the document review experience: PDF and image rendering, page navigation, bounding-box overlays, confidence highlighting, side-by-side source-to-extraction linking
- Build human-in-the-loop correction workflows — fast keyboard-driven review, field-level edit, approve/reject queues — and wire corrections back into the evaluation loop
- Build the extraction schema designer: let users define, test, and version the structured output they want from a document type
- Build internal evaluation and observability dashboards — accuracy by field and document type, latency and cost breakdowns, model comparison views Handle real-time and long-running state well: streaming partial results, job progress, optimistic updates, resumable sessions
- Own frontend performance on genuinely heavy documents — virtualisation, canvas rendering, lazy loading, memory discipline
- Partner with backend engineers on API contracts; you will shape them, not just consume them
- Set the frontend standard for the team: component architecture, typing, testing, accessibility
What We're Looking For
- 4–5 years building production frontend applications, with strong ownership of what you shipped
- Deep React expertise — hooks, rendering behaviour, state management at scale, performance profiling; you know why a component re-renders
- Strong TypeScript; you use the type system to make bad states unrepresentable Real command of the browser platform: layout, events, memory, rendering; comfort with Canvas or SVG for custom rendering
- Experience with complex, data-dense product surfaces — dashboards, editors, annotation or review tools — rather than marketing pages
- Solid API integration instincts: async state, caching, polling and streaming, error and retry handling
- Design sensibility; you can take a rough direction and ship something clean without a pixel-perfect mock
Bonus Points
- Experience with PDF rendering in the browser (PDF.js or equivalent) or with annotation/labelling tools
- Familiarity with WebSockets, SSE, or other streaming transports
- Exposure to AI/ML product surfaces — human-in-the-loop review, model output inspection, eval tooling
- Comfort dropping into a Python or Go backend when it unblocks you Design systems ownership, or strong accessibility practice
Note
We are looking for people who can own the outcomes described here, not people who match every line of this specification. If this problem excites you and you believe you can do this work, we want to hear from you.
Why Sarvam?
Sarvam is a fast-moving, high talent-density team building full-stack AI for India, working on problems that push the frontiers ofAI with real population-scale impact.
Work alongside researchers, engineers, builders, and business leaders who move fast and hold each other to a very high bar
High ownership and high impact, from day one
Everything we do is AI-first, from the way we build and ship to the way we think about problems
You can work on problems that could change how an entire country learns, works, and communicates
If you want to work on problems at the frontier ofAI in India, Sarvam is the place to be.
Click on Apply to know more.
