SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Sarvam AI is India's leading full-stack AI company building sovereign AI infrastructure and applications for the Indian market. This role is part of a small team building Sarvam's context engine—a platform that enables enterprise AI agents to work with real data across documents, databases, and business systems, while maintaining memory of what they learn over time.
You will own large parts of the platform, working on systems that directly impact customer deployments. The work spans the full stack: building ingestion pipelines that parse, extract, and index large document volumes reliably; connecting to databases and enterprise systems for structured data access; designing how enterprise data is modeled including ontologies and entity matching; building retrieval tools with search, filters, and citations; implementing agent memory at user, team, and organizational levels; running quality evaluations and iterating on results; and collaborating with product and forward-deployed engineers to scale customer-specific solutions across the platform.
The production stack is Go, Temporal, OpenSearch, and Postgres. You will write design docs, keep decisions documented, and work closely with teams building Sarvam's models, document AI, and inference systems.
REQUIREMENTS:
- 3–5 years of experience building and running backend or data systems in production
- Strong fundamentals in at least one backend language; Go experience is a plus and you should be comfortable working in it daily
- Solid data modeling and PostgreSQL knowledge
- Experience with long-running pipelines, retry logic, and recovery patterns
- Experience with search or retrieval systems and debugging retrieval quality
- Daily use of AI coding tools with good judgment about their strengths and limitations
- Comfort measuring quality and building evaluation frameworks
BONUS: Knowledge graphs, ontologies, entity resolution, structured extraction, agent memory for LLM applications, document processing (OCR, layout, table extraction), ERP/CRM connectors or change data capture, MCP, serving LLMs and embeddings in production, on-prem or air-gapped deployments, early- or growth-stage startup experience.
