2 days ago
Base Salary
$143k - $291k/yr
Responsibilities
- Investigate, design, develop, execute, and implement AI tokenomics research for new products, technologies, and intellectual property.
- Research and prototype methods for representing jobs and system states and predicting workload complexity, answer quality, latency, token demand, energy, and cost.
- Develop routing policies, model cascades, adaptive controllers, and budget-allocation methods for LLMs, reasoning modes, retrieval strategies, tool chains, and deployment targets.
- Build benchmark, telemetry, accounting, and workload-simulation environments.
- Conduct controlled experiments, ablation studies, sensitivity analyses, and stress tests across models, hardware platforms, and deployment environments.
- Evaluate answer quality, routing accuracy, token efficiency, tail latency, throughput, reliability, cost, energy, and carbon metrics.
- Collaborate with internal and external researchers and engineers to validate results and transfer technologies into HPE AI platforms.
- Prepare research papers, presentations, patent disclosures, and supporting documentation.
Requirements
- Completed or soon-to-be-completed PhD in Computer Science, Electrical Engineering, Statistics, Applied Mathematics, or a related field, with a machine-learning research focus.
- Typically four years of experience, including graduate and doctoral research.
- Strong foundation in deep learning, optimization, probabilistic modeling, reinforcement learning, or related machine-learning methods, demonstrated through doctoral research, internships, or publications.
- Ability to implement and evaluate research ideas in Python using frameworks such as PyTorch, with sound experimental and software-engineering practices.
- Experience applying experimental design, benchmarking, ablation studies, and statistical analysis to research data.
- Strong analytical, problem-solving, written, and verbal communication skills.
- Ability to communicate research plans, proposals, and results and negotiate options at management levels.
- Preferred experience with ML, LLMs, model routing, reinforcement learning, causal inference, constrained optimization, retrieval-augmented generation, or agentic AI.
- Preferred interest or experience in token-efficient inference, per-job compute budgeting, multi-model serving, edge-to-cloud AI systems, sustainable AI, energy-aware computing, or carbon-aware scheduling and optimization.
Benefits
- Hybrid work arrangement requiring an average of two days per week from an HPE office.
- Health and wellbeing benefits supporting physical, financial, and emotional wellbeing.
- Personal and professional development programs supporting career growth and subject-matter expertise.
- Inclusive workplace culture and accessibility accommodations for qualified applicants.
Categories
AI ResearchML Engineering
About HPE
Hewlett Packard Enterprise (HPE) builds enterprise IT infrastructure and services, including servers, storage, networking (Aruba), and hybrid edge-to-cloud platforms like HPE GreenLake for businesses and public-sector organizations. Formed in 2015 after Hewlett-Packard’s split, HPE is headquartered in Houston, Texas and trades on the NYSE. It sells hardware, software, and subscription-based managed cloud services to help customers run workloads across data centers, colocation, and public clouds.