Vice President of Research, Machine Learning

About A1

A1 serves over 5 billion users with fundamental applications such as email, notes, and task management, which lack integrated AI capabilities. Our mission is to build a proactive smart assistant that infuses intelligence into daily conversations, errands, organization, and workflows with minimal user prompting. Our product emphasizes exceptional reliability in managing extended workflows with persistent context, real-world task completion, multi-step reasoning, interaction with external tools, and consistent behavior despite unpredictable models. The goal is to significantly reduce daily task time by approximately 90% while enhancing user enjoyment.

Role Overview

As the VP of Research for Machine Learning, you will be responsible for steering the research and intelligence strategies that power our core product. You will define how artificial intelligence reasons, assesses, and advances within a frequently used consumer product.

Key Responsibilities

• Establish and refine research focus areas such as context representation, memory management, reasoning, planning, and orchestration within A1's intelligent systems.
• Evaluate and decide between creating novel model architectures or utilizing leading-edge open-source and commercial models.
• Develop comprehensive evaluation methodologies prioritizing practical usefulness, robustness, safety, and sustained performance over transient benchmark results.
• Lead efforts in alignment, safety, and guardrail strategies, treating them as integral product aspects.
• Explore advanced techniques including retrieval-augmented training, mixture-of-experts models, knowledge distillation, multi-agent orchestration, and multimodal system integration.
• Collaborate closely with product teams and application engineers to shape early-stage product intelligence.
• Set and uphold high standards for research rigor, discerning judgment, and aesthetic consideration throughout the organization.

Qualifications

• Extensive experience developing or enhancing operational machine learning systems deployed in production environments.
• Strong technical insight into model behaviors, potential failure modes, and decisions involving long-term trade-offs.
• A pragmatic approach focused on building effective real-world systems rather than solely theoretical ideas.
• Capability to make significant, sometimes irreversible decisions amid incomplete information.
• A rigorous and persistent focus on evaluation accuracy, correctness, and system behavior over extended durations.
• High level of ownership and entrepreneurial mindset, working more like a founder than a conventional manager.
• Note: Candidates mainly interested in publishing, incremental benchmarks, or managing large research groups may find this role unsuitable.

Technical Environment

• Programming in Python.
• Experience with PyTorch and JAX frameworks.
• Utilization of GPU-based platforms for training and inference tasks.

Work Culture and Process

We believe top-tier products emerge from compact, exceptional teams with high talent density operating hands-on. We embrace rapid decision-making with collective input, balancing swift delivery with continuous learning. The role demands structured thinking, sound judgment, and independent execution to deliver a truly transformative product to our users.

Interview Procedure

Qualified candidates will undergo 3 to 4 interviews conducted virtually or onsite with technical team members. We prioritize transparency and efficiency, providing prompt decisions. Exceptional applicants will receive offers to join a team committed to delivering practical AI benefits to billions worldwide.