Granica
Research Product Manager – AI Systems
3mo ago
160000 –250000 USD / yearUSAMiddleml infrastructuremodel evaluationbenchmarkingstructured dataperformance optimization
Manage product development for AI systems focusing on model evaluation, continuous learning, and performance optimization.
Responsibilities
- We’re hiring a Research Product Manager to define and build core systems that determine how AI models are evaluated, improved, and deployed on real-world data.
- You’ll work on systems spanning:
- model evaluation and benchmarking
- post-training and feedback loops
- structured and relational data learning
- performance, efficiency, and cost optimization
- This role sits at the intersection of ML infrastructure, research, and product . It is closest to roles like ML platform PM or AI infrastructure PM , but with deeper ownership of how systems are designed and how model performance translates into real-world outcomes.
- You’ll partner closely with researchers and engineers to move ideas from experiments into production systems used at scale.
Requirements
- 5+ years of experience in product management, technical program management, or similar roles in AI, ML infrastructure, or data systems
- Strong understanding of machine learning systems , including training, evaluation, and deployment
- Experience working with large-scale data systems or distributed infrastructure
- Ability to reason about trade-offs across data, compute, performance, and cost
- Track record of driving complex technical systems from concept to production
Nice to have
- Experience with ML platforms, LLM systems, or AI infrastructure
- Experience with evaluation systems, observability, or model performance tooling
- Familiarity with structured or relational data systems (e.g., warehouses, lakehouses)
- Background in engineering, applied research, or ML systems development
- Experience operating in research-driven or highly ambiguous environments
Conditions
- Competitive salary, meaningful equity, and performance bonus for top performers
- 401(k) with company match, comprehensive health coverage, and unlimited PTO
- Daily catered meals in our Mountain View office
- Support for research, publication, and conference participation
- At Granica, you'll help build the next generation of enterprise AI —from exabyte-scale data infrastructure , Large Tabular Models (LTMs) , and stateful AI agents . Together, we're creating the infrastructure that enables enterprises to own their data , own the intelligence built on it , and scale both efficiently .
Other
- AI today is no longer bottlenecked by model architecture alone.
- The real constraints are:
- how models are evaluated
- how they improve after training
- how they behave in real-world systems
- Granica is building the systems that solve this.
- We are a research and systems company led by Prof. Andrea Montanari (Stanford) , focused on:
- evaluation as a first-class system
- post-training as a continuous learning loop
- efficient learning over real-world data
- Most real-world data is structured and relational , yet modern AI systems remain poorly optimized to learn from it.
- Our thesis: AI advantage will come from how efficiently models learn from structured data—and how that translates into economic value.
- Define and drive systems for model evaluation, benchmarking, and real-world performance
- Build product direction for post-training systems and feedback loops that continuously improve models
- Define how models learn from large-scale structured and relational datasets
- Partner with engineering to build systems that connect data platforms (warehouses, lakehouses) with ML systems
- Own how improvements move from research experiments into production systems
- Model trade-offs across compute, data efficiency, performance, and cost
- Identify where system improvements drive measurable business impact
- ML / AI infrastructure PMs (OpenAI, Google, Meta, Snowflake, Databricks, AWS, or similar)
- Product leaders in model systems, evaluation, or observability
- Research engineers or applied scientists transitioning into product
- Engineers who have built ML or data systems and taken on product ownership
- Most AI systems are limited not by model capability, but by:
- weak evaluation systems
- inefficient learning loops
- poor utilization of structured data
- lack of connection between performance and real-world outcomes
- This role defines how those constraints are solved in production systems.
- You won’t be optimizing features—you’ll be defining the systems that determine how models improve, how they are trusted, and how they deliver value.
- Location: Mountain View, CA
- Work model: On-site, five days per week
- Level: Senior / Staff / Principal (depending on experience)