Software Development Engineer - Tejas
Salary not listed • Bengaluru • Full-time
Posted 8d ago
Remote Work Policy
On-site
Employment Type
Full-time
Categories
AI Agent Engineer
About the job
Our team builds Sarvam's context engine to solve key challenges for AI agents interacting with enterprise data. We aim to enable agents to effectively search, navigate, and cite information from diverse sources like documents, databases, and business systems, overcoming the limitations of current agents that forget information after a session. We are developing both a knowledge system to consolidate enterprise data and a memory system for agents to retain learning over time, focusing on continuous improvement and measurable results that outperform traditional approaches. This role offers the opportunity to own significant parts of the platform, with work directly driven by customer deployments, allowing for rapid iteration and close collaboration with teams working on Sarvam's models, document AI, and inference.
Responsibilities
- Build and maintain the ingestion pipeline for parsing, extracting, and indexing large volumes of documents, ensuring reliability even with messy files and failing external services.
- Integrate with databases and enterprise systems to enable agents to work with both structured records and documents.
- Design the data modeling for enterprise data, including ontology and entity matching across sources, with traceability for every value.
- Develop retrieval tools for agents, including search, filters, and citations, optimizing for speed and cost-effectiveness.
- Build and manage agent memory at user, team, and organizational levels, focusing on storage, retrieval, and accuracy over time.
- Conduct evaluations (evals) to identify and rectify answer inaccuracies, driving continuous quality improvement.
- Collaborate with product and forward-deployed engineers to translate customer needs into platform-wide features.
- Create and maintain design documentation for all decisions.
Requirements
- 3 to 5 years of experience in building and running backend or data systems in production, with a proven ability to troubleshoot and fix issues.
- Strong proficiency in at least one backend language; Go experience is a plus, with a willingness to work in Go daily.
- Solid fundamentals in data modeling and PostgreSQL.
- Experience with long-running pipelines, including ensuring retry and recovery mechanisms.
- Experience with search or retrieval systems, including debugging result discrepancies.
- Proficiency in using AI coding tools and understanding their practical applications and limitations.
- Commitment to measuring quality and experience in building evaluation frameworks.