Back
Strong Tier-1 brand, metro Bengaluru, and a Product Manager title create high competition despite niche ML infrastructure focus.
Deep LLM infrastructure, capacity modeling, and vendor management needs make the role highly domain-specific.
Explicit 10–12 years requirement plus mandatory coding, LLM infrastructure, and vendor experience makes filters highly strict.
Login to See Your Match Score
Create a free account or log in to unlock your CV match score across:
Own and operate automated systems monitoring TPM/RPM utilization across a global LLM fleet covering multiple models, regions, and providers in real time to preempt incidents.
Forecast and model capacity demand for trillion-token workloads to inform hyperscaler commitments and cluster sizing while balancing latency SLAs and cost.
Automate capacity management via scripting and pipelines to handle reservations, rate limiting, failover, and dynamic spend optimization across 1st party and 3rd party providers.
10 to 12 years of experience in technical or infrastructure product management or equivalent role involving hands-on building, not just specification.
Proficiency in Python programming with ability to write efficient automation scripts and leverage coding AI agents for operational tasks.
Strong data and mathematical fluency, including ability to extract and analyze telemetry data using SQL or Splunk and build forecasting models.
Deep knowledge of LLM infrastructure concepts including token rate limits, PTU versus PayGo economics, model latency, and capacity management across proprietary and API-based models.
Experienced in managing large-scale AI infrastructure operationally with capability to independently build tooling and automation rather than relying on engineering resources.
Familiar with multi-provider LLM routing and capacity management tradeoffs on hyperscalers such as Azure OpenAI, Google Vertex, or AWS Bedrock.
Skilled in vendor management for 3rd party LLM providers, translating technical usage data into smart procurement and operational decisions.