Jobs at Sarvam

Own product roadmaps for ASR, TTS, and model evaluation. Coordinate data collection, training cycles, and releases. Build evaluation frameworks and run experiments (offline, A/B, human) to improve accuracy, latency, MOS, and robustness. Ensure scalable deployment, monitoring, and translate business needs into measurable model goals.

Deploy alongside clients to transform heterogeneous multimodal data into semantic structures for AI reasoning. Design ontologies, segmentation, embedding and retrieval strategies, collaborate on ingestion pipelines, evaluate retrieval and reasoning performance, and translate insights into product and engineering signals.

Design and maintain domain-specific evaluation frameworks for LLMs and AI systems: define quality metrics, run pre/post-deployment evaluation cycles, build dashboards, find failure modes and distribution shifts, operationalize automated eval pipelines with MLOps, manage datasets and annotation workflows, and publish quality reports.

Design, build, and maintain scalable backend services and APIs for a voice/AI dialogue platform. Optimize schemas and queries, ensure observability and reliability, resolve scaling/latency issues, integrate models into production, enforce security and best practices, and collaborate across product, frontend, and DevOps teams.

Lead go-to-market and positioning for enterprise AI products. Build vertical narratives, sales enablement content, case studies, demos, and ROI benchmarks. Partner with product and sales, gather field insights, and support long-cycle enterprise deals in financial services, public sector, and healthcare.

Own and scale partnerships with India-based and India-operating global GSIs. Drive joint business plans, co-sell motions, partner-sourced deals, co-marketing, certification programs, and product feedback to generate measurable revenue across BFSI, public sector, manufacturing, and telco.

Build and scale high-throughput, low-latency Vision APIs and platform infrastructure. Develop Python/FastAPI services, implement observability, security, CI/CD, container orchestration, async task processing, and real-time HTTP/WebSocket communication across cloud environments.

Conduct research across vision-language model lifecycles: architecture, training (pretraining, SFT, RLHF, DPO), data strategies, evaluation and benchmarks for Indic multimodal tasks; diagnose failures, improve robustness and interpretability, prototype at scale, and engage with open-source and research community.

Own the full enterprise sales cycle: prospect, engage C-level stakeholders, conduct discovery, negotiate and close deals. Collaborate with BDRs/SDRs, marketing, and product, maintain CRM discipline, forecast pipeline, and ensure smooth handoffs to customer success to drive acquisition and long-term partnerships.

Drive outbound prospecting and lead qualification via cold calls, email, and LinkedIn; engage enterprise decision-makers, conduct discovery calls, maintain CRM records, and collaborate with AEs and marketing to grow pipeline.

Build distributed microservices and task-graph execution engines for LLM-driven workflows. Integrate multiple LLM providers and tools, implement gRPC/REST APIs, document ingestion and embedding pipelines, vector search knowledge bases, safe code-execution sandboxes, and contribute to MCP server implementations while maintaining containerized infrastructure and observability.

Design, build, and maintain ML systems for a multilingual dubbing platform. Develop automated validation, monitoring, and evaluation workflows. Optimize inference pipelines for performance, cost, and reliability, run experiments, and collaborate with product and engineering to productionize models.

Senior client-facing role owning GTM, pipeline, and account strategy for defence and strategic clients. Build decision-level relationships, translate technical AI capability into proposals, lead RFI/RFP and PoC scoping, negotiate commercially, collaborate with engineers and product, manage account health, and surface requirements and competitor/policy shifts.

Own frontend quality across web, streaming, and AI interaction layers. Define and implement testing strategy, build automation systems (component/integration/visual regression), use AI to accelerate testing, debug browsers, and partner with frontend engineers to prevent incidents and block releases when quality thresholds arent met.

Own the product roadmap for Chanakya atoms, translate field deployments into precise specs, run structured enterprise discovery, prioritise features, define product metrics, write PRDs and decision logs, coordinate engineering sprints, and manage stakeholder communication to deliver deployable AI components.

Drive end-to-end enterprise sales for Sarvam's conversational AI solutions to large BFSI organisations. Build trusted relationships with decision-makers, understand client pain points, coordinate internal teams to structure solutions, manage complex procurement and multi-stakeholder deals, ensure smooth deployment with customer success, and monitor market trends and competitive landscape.

Lead backend engineering for Document Intelligence and Vision APIs: architect for low latency and high throughput, scale services, build observability and security foundations, manage infra and deployments, improve developer experience, mentor engineers, and define CI/CD and inference-serving strategies.

Operate and own model lifecycle for defence and strategic deployments: design serving infra, CI/CD for model updates, monitoring/observability, evaluation/A-B testing, containerised serving for edge and air-gapped environments, collaborate on eval pipelines, create runbooks, and lead incident response for production model failures.

Design and implement production-grade backend systems for on-prem and constrained deployments: MCP servers, ingestion pipelines, agentic backends, APIs, containerised tooling, and operational/debugging support for deployments.

Build and optimize scalable Next.js and TypeScript web applications, create reusable components, integrate RESTful APIs, ensure performance, accessibility, and cross-browser compatibility, maintain quality through testing and code reviews, and collaborate with designers, backend engineers, and AI teams using monitoring, analytics, and A/B testing.