Staff AI Engineer

Teradata Group; Hyderabad; Presencial; INR 4 000 000 - 7 000 000; Tempo integral

Resumo da oferta

Teradata is seeking a Staff AI Engineer in Hyderabad to shape the next generation of AI-native analytics. You’ll own the architecture of a secure, multi-tenant agentic AI platform, integrating LLMs and Tooling, and building ModelOps pipelines for production reliability.

You’ll lead cross-team design reviews and drive scalable features across Teradata’s cloud-native SaaS stack, while enforcing guardrails and cost controls for enterprise workloads.

Qualificações

• BS/MS/PhD in Computer Science, AI/ML, or related field.
• 8+ years of professional software engineering experience with ownership of complex production systems.
• 2+ years building LLM-based applications or platforms in production.
• Strong cloud engineering on AWS, Azure or Google Cloud: Kubernetes, containers, IaC (e.g. Terraform), CI/CD.
• Experience building SaaS with multi-tenancy, APIs, authentication/authorization, observability and SLOs.
• Hands-on LLM experience: prompt and context engineering, function and tool calling.
• Experience designing agentic systems: planning, memory, orchestration; knowledge of LangGraph/LangChain.
• ModelOps/LLMOps: deploying, versioning, evaluation and monitoring models, prompts and agents.
• Python (or Java/Go) production-grade systems.
• Experience with large-scale data and distributed systems.

Responsabilidades

• Own architecture and technical direction of Teradata’s agentic AI platform as a secure, multi-tenant SaaS service.
• Design and build agent runtimes and harnesses: tool calling, planning, memory, context management, orchestration and sandboxed execution.
• Design, build and secure MCP servers/clients and agent APIs.
• Integrate and serve hosted and self-hosted LLMs: model gateway, routing, caching, latency and cost control.
• Build ModelOps/LLMOps pipelines: registries, CI/CD with gates, canary releases, rollback and monitoring.
• Develop evaluation and observability: offline evals, tracing, latency, cost metrics and SLA measurement.
• Implement AI guardrails: prompt-injection defense, PII handling, policy enforcement and audit trails.
• Run SaaS AI services on AWS/Azure/Google Cloud: tenant isolation, quotas, SLOs and on-call readiness.
• Improve scalability and cost efficiency using Kubernetes and IaC; apply RAG and vector search where relevant.
• Lead design reviews; collaborate with product, research and platform teams to scale customer use cases.

Conhecimentos

Python

Java

Go

Kubernetes

Terraform

OAuth/OIDC

Formação académica

BS/MS/PhD in Computer Science, AI/ML or related field

Ferramentas

Kubernetes

Terraform

LangChain

LangGraph

OAuth/OIDC

Descrição da oferta de emprego

Our Company:

Teradata is building the next generation of AI-native analytics, enabling customers to deploy production-grade Generative AI systems directly where enterprise data lives. We are looking for a Staff AI Engineer to play a key role in designing and building the LLM, agent and ModelOps capabilities of Teradata’s cloud and SaaS platform, taking them from prototype to secure, observable, multi-tenant production services.

This role is ideal for an engineer who thrives at the intersection of LLMs, agentic systems and cloud-native engineering, and wants to work on core platform capabilities, not just application demos.

What You’ll Do: Shape the Way the World Understands Data

You will:

• Own the architecture and technical direction of Teradata’s agentic AI platform, delivered as a secure, multi-tenant, cloud-native SaaS service.

• Design and build agent runtimes and harnesses: tool calling, planning, memory, context management, multi-agent orchestration, sandboxed execution and human-in-the-loop controls.

• Design, build and secure MCP servers and clients, and other agent and tool protocols, that expose Teradata capabilities to agents through governed, well-documented APIs.

• Integrate and serve LLMs, both hosted APIs and self-hosted open-weight models: model gateway, routing, caching, fallbacks, token budgets, latency and cost control.

• Build ModelOps / LLMOps pipelines: model, prompt and agent registries and versioning, CI/CD with evaluation gates, canary releases and rollback, and quality and drift monitoring.

• Build evaluation and observability frameworks for LLM and agent systems: offline evals, LLM-as-judge, tracing, regression suites, and quality, safety, latency and cost metrics.

• Implement AI guardrails and responsible-AI controls: prompt-injection defense, PII handling, permissions, policy enforcement and audit trails for regulated industries.

• Run AI services as SaaS on AWS, Azure and Google Cloud: tenant isolation, authentication and authorization, quotas, usage metering, SLOs, and production and on-call readiness.

• Drive scalability, performance, reliability and cost efficiency of AI workloads, using Kubernetes and infrastructure-as-code.

• Where use cases call for it, apply RAG, vector search and semantic retrieval patterns, including Teradata’s vector store.

• Lead technical design reviews and influence cross-team architectural decisions; identify risks early and drive solutions.

• Collaborate with product, research and platform teams to turn customer use cases into scalable features, and benchmark against cloud and open-source alternatives.

Who You’ll Work With: Join Forces with the Best

You’ll collaborate with a world-class team of AI architects, ML engineers and domain experts across India and Silicon Valley, working together to build the next generation of enterprise AI systems.

You’ll also work cross-functionally with:

• Product managers and UX designers to craft agentic workflows that are intuitive and impactful.

• Domain specialists to ensure solutions fit real business problems in regulated industries.

• Cloud, SaaS and infrastructure teams responsible for deploying, evaluating and scaling AI workloads.

This is a rare opportunity to shape foundational AI capabilities within a global, data-driven company. This is a deeply collaborative environment where technical innovation meets real-world application, where your ideas are not only heard but implemented to shape the next generation of data interaction.

Minimum Requirements

• BS/MS/PhD in Computer Science, AI/ML, or a related field.

• 8+ years of professional software engineering experience, including ownership of complex production systems, and 2+ years building LLM-based applications or platforms in production.

• Strong cloud engineering on AWS, Azure or Google Cloud: Kubernetes and containers, infrastructure-as-code (for example Terraform), and CI/CD.

• Experience building and operating SaaS services: multi-tenancy, API design, authentication and authorization (OAuth/OIDC), observability, SLOs and reliability engineering.

• Hands-on LLM experience: prompt and context engineering, function and tool calling, structured outputs, and working with both commercial and open-weight models.

• Experience designing agentic systems: planning, memory, orchestration and custom agent harnesses, plus familiarity with agent frameworks such as LangGraph, LangChain or equivalent.

• Experience with ModelOps / LLMOps: deploying, versioning, evaluating and monitoring models, prompts and agents in production.

• Working knowledge of methods for evaluating LLM and agent quality, safety, latency and cost.

• Proficiency in Python (and/or Java or Go), with experience shipping production-grade systems.

• Experience with large-scale data and performance-sensitive distributed systems.

Preferred Qualifications

• Experience with LLM inference and serving optimization (for example vLLM or Triton, quantization, GPU scheduling).

• Experience with fine-tuning and post-training of open-weight models (for example LoRA, preference optimization, distillation).

• Experience implementing AI guardrails, red-teaming and security controls for agents, such as prompt-injection defense.

• Familiarity with agent-to-agent protocols and tracing standards such as A2A and OpenTelemetry for GenAI.

• Experience with analytics platforms, data warehouses and SQL-based workloads.

#LI-PG1