SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Sarvam AI is building India's sovereign AI platform, developing full-stack AI infrastructure with a focus on making AI work for India at scale. The company is backed by Lightspeed, Peak XV, and Khosla Ventures, and partners with leading Indian enterprises including Tata Capital, SBI Life, CRED, IDFC, and LIC.
The Vision team builds research-grade vision-language models for OCR and structured document extraction, then engineers the production serving harness that transforms these models into a document intelligence platform. The engineering challenge is to match frontier models like Gemini Flash using smaller sovereign models, at a fraction of the cost and fully within India.
As Frontend Engineer for Vision, you will own the user-facing interfaces that make document intelligence trustworthy and actionable. You'll build document viewers with field-level grounding and bounding-box overlays, human-in-the-loop correction workflows with fast keyboard-driven review, extraction schema builders, evaluation dashboards, and the developer console for enterprise customers running pipelines at scale. These are dense, stateful, performance-sensitive UIs handling 100-page PDFs with streaming results and real-time corrections.
You will be the frontend owner for the team, working closely with backend and applied AI engineers rather than against a finished spec. You'll own frontend performance on genuinely heavy documents through virtualization, canvas rendering, lazy loading, and memory discipline. You'll partner with backend engineers on API contracts, shaping them rather than just consuming them, and set the frontend standard for the team in component architecture, typing, testing, and accessibility.
The stack includes Go, Python, Temporal, REST, Kubernetes, PostgreSQL, Redis, object storage, and OpenTelemetry-based observability. You'll work on the full messiness of Indian documents at population scale: PAN and Aadhaar, bank statements, GST filings, insurance and medical reports, contracts, legal filings—across languages, scan quality, and layouts never designed to be machine-read.
