AI Infrastructure / DevOps Engineer (AI-Native, Agentic) 5–10 Years

Sprouts.ai; Chandigarh; Hybrid; INR 1,200,000 - 1,800,000; Full time

Benefits offered by this job

Ownership of infra decisions

Fast track to leadership roles

Exposure to frontier AI systems

Job summary

A leading AI technology company in Chandigarh is seeking a Foundational AI Infrastructure Engineer to oversee the architecture and deployment of AI-native infrastructures. The role demands expertise in Kubernetes and Terraform, with responsibilities ranging from optimizing GPU workflows to implementing observability measures. Ideal candidates will have hands-on experience with model-serving infrastructure and a proactive approach to problem-solving in high-speed environments. This is an opportunity to shape the future of AI-driven platforms.

Qualifications

• Strong production experience with Kubernetes + Terraform.
• Hands-on experience with GPU and model-serving infrastructure.
• Ability to read and understand Python/Node code.

Responsibilities

• Architect and own the AI-native infrastructure.
• Deploy and optimize GPU and LLM runtimes.
• Implement event-driven architectures.

Skills

Kubernetes + Terraform

Cloud infra ownership (AWS/GCP/Azure)

GPU or model-serving infrastructure

Vector databases or high-performance search

Read Python/Node code

Optimizing systems for cost + latency

Real observability implementation

Tools

Terraform

Docker

OpenTelemetry

Job description

Job Type: Full-Time (Hybrid Mode)

Sprouts.AI is an AI-native “Generative Demand Platform” disrupting the $1.1T B2B GTM category. We help demand-gen and sales teams generate qualified pipelines through real-time intelligence signals and agentic AI execution. Customers are already seeing:

• 50%+ increase in SDR/BDR productivity
• 30%+ reduction in martech spend

As we scale, we’re building the next-generation infrastructure for massive data ingestion, vector search, multi-agent orchestration, real-time LLM workloads, and GPU-powered inference. We’re hiring a Foundational AI Infrastructure Engineer who can architect and operate this system end-to-end.

This is not a typical DevOps role.

You will not “wait for requirements.”

You will read the code, understand the system, and shape the infra proactively.

If you want to build the backbone of a high-velocity, agentic AI platform, this is your role.

What You’ll Own (Outcomes, Not Tasks)1. AI-Native Infrastructure Ownership

• Architect and own the compute, storage, networking, vector search, orchestration, and inference layers.
• Derive infra needs directly from backend + AI pipelines.
• Build infra that supports massive concurrency, low latency, and unpredictable agent workflows.

2. GPU & LLM Inference Readiness

• Deploy and optimize LLM runtimes (vLLM, TGI, TensorRT-LLM).
• Configure GPU autoscaling, batching, quantization, caching.
• Ensure multi-cloud portability and avoid vendor lock-in.

3. Event-Driven Architecture for Agentic Systems

• Implement backpressure, retries, DLQs, and idempotent pipelines.
• Support parallel agent execution and high-throughput ingestion.

4. High-Performance API Gateway & Traffic Control

• Implement rate limits, quotas, throttling, and tenant-isolated traffic shaping.
• Prevent token floods and agent-triggered burst loads.

5. Internal Developer Platform (IDP)

• Build reusable infra templates, CI/CD scaffolds, local dev reproducibility, canary deployments, rollback mechanisms.
• Enable engineers to ship in hours, not weeks.

6. Observability “Everywhere”

• LLM-level observability (token usage, eval metrics, latency maps).
• Build dashboards for agent behavior, vector search performance, API latency, GPU utilization.
• Build transparent cost dashboards for compute, storage, token usage, and inference.
• Optimize infra for cost-performance (batching, caching, autoscaling, routing).
• Keep infra costs predictable as LLM usage scales.

8. Data Lifecycle & Index Management

• Define TTL, retention, compaction, cold storage rules.
• Ensure metadata consistency at scale.

9. Security, Governance & Multi-Tenancy

• Ensure safe tool execution for agents (sandboxing, ACLs, rate limits).
• Zero Trust patterns for internal services.

10. Code-Level Performance Insights

• Read Python/Node code to identify bottlenecks, parallelism issues, inefficient queries, unbounded concurrency, missing caching.
• Suggest backend and AI pipeline optimizations that reduce infra load.

Tech You’ll Work WithCloud, Compute & Orchestration

Terraform (must)

Docker

Load balancers, service mesh

AI/LLM Infrastructure

vLLM, TGI, TensorRT-LLM

Batching, quantization, caching

Token routing + observability

Data & Vector Search

Monitoring & Reliability

OpenTelemetry traces

Loki / ELK stacks

P95/P99 latency optimization

GitHub Actions / GitLab

GitOps (ArgoCD optional)

Infra templates, one-click deploys

Must-Have Skills (Hard Filters)

• Strong production experience with Kubernetes + Terraform
• Cloud infra ownership (AWS/GCP/Azure)
• Experience with GPU or model-serving infra (even small-scale)
• Hands-on with vector DBs or high-performance search
• Ability to read Python/Node code and infer infra impact
• Proven history of optimizing complex systems for cost + latency
• Real observability implementation experience (metrics, traces, logs)

Nice-to-Haves (Signal Boosters)

• vLLM or TGI deployment experience
• Knowledge of LangGraph/LangChain internals
• Experience building IDP / self-serve platforms
• Advanced GPU optimization experience
• Built multi-region or multi-cloud systems
• Experience with OpenTelemetry for multi-agent architectures
• Hands-on RAG or AI pipeline infra work
• Optimization of P95/P99 latencies

You’ll Thrive Here If You…

• Prefer proactive problem discovery over reactive “fix requests.”
• Can zoom between 5,000-foot architecture and 5-line code-level insights.
• Love squeezing milliseconds out of systems.
• Believe infra should accelerate product velocity, not slow it down.
• Enjoy the chaos, creativity, and pace of a high-speed AI-native company.

Please Don’t Apply If…

• You want a traditional DevOps job focusing only on deployments and clusters.
• You cannot read backend code or understand LLM/agent workflows.
• You rely too heavily on vendor-managed AI services.
• You need strict processes and long planning cycles.
• You’ve owned infra for a high-scale or AI-heavy system.
• You’ve debugged distributed systems in production.
• You can explain cost vs. latency trade-offs clearly.
• You’ve tuned vector DB search, async workers, or GPU workloads.
• You’ve built observability or IDP tooling used by other engineers.
• You have a “this is inefficient” instinct just by scanning a diagram or code snippet.

What You’ll Get

• A mission-critical role shaping the foundation of Sprouts.ai’s AI-native platform.
• Ownership of infra decisions across compute, data, storage, AI inference, observability, and cost.
• The opportunity to architect GPU-ready infra for future private LLM deployments.
• Fast track to Head of Infra / Platform for strong performers.
• A culture of rapid iteration, autonomy, and engineering excellence.
• Deep exposure to frontier AI systems, multi-agent orchestration, and large-scale vector search.