AI Scientist — Commerce Labs

Salesforce•Burlington, MA

14d•$117,200 - $246,000•Hybrid

About The Position

Salesforce is the #1 AI CRM, where humans with agents drive customer success together. The Experience Commerce Labs is where Salesforce works on the future of commerce. We build agentic systems that learn efficiently over time from real interactions with customers, merchants, and the marketplace environment, pushing the boundary of what autonomous commerce agents can do at enterprise scale. Commerce Labs builds on top of Salesforce Commerce Cloud, Salesforce's unified commerce platform that powers storefronts, checkout, order management, and merchandising for enterprise brands and retailers around the world. Commerce Cloud handles massive scale and complexity, including high-traffic storefronts, peak shopping events, multi-channel selling, and deep integration with the broader Salesforce Customer 360 ecosystem (Service, Marketing, Data Cloud, and more). This gives Commerce Labs a unique advantage: our agentic research isn't theoretical, it's grounded in real production traffic, real merchant data, and real customer behavior at scale, so what we build can go straight into the hands of the retailers who run their businesses on Commerce Cloud. We're hiring an AI Scientist to design, prototype, and ship machine learning (ML) systems that power our agentic commerce products. You'll work closely with research and product engineering to take ideas from whiteboard to production, with real customers and real-world data, and with direct line of sight into how your work shows up in Commerce Cloud storefronts.

Requirements

You have hands-on machine learning experience, meaning you've built, trained, and evaluated ML models yourself, not just directed others.
You have a strong foundation in one or more of: reinforcement learning (RL), deep learning, optimization, causal inference, or operations research.
You're proficient in Python and standard ML tooling (for example, PyTorch, TensorFlow, JAX).
You have strong problem-solving skills and comfort working with ambiguous, open-ended problems.
You're comfortable moving fast from idea to working prototype.

Nice To Haves

You have a PhD (completed or in progress) in Applied Science, Machine Learning, Electrical Engineering, Computer Science, Operations Research, or a related field.
You have a strong record of published basic or applied AI research, for example at top-tier conferences or journals such as NeurIPS, ICML, ICLR, AAAI, CVPR, RecSys, EC, EMNLP, ACL, or equivalent.
You have experience with agentic or multi-step decision-making systems, such as RL agents, large language model (LLM) based agents, or sequential decision-making.
You have prior experience applying ML to commerce, marketplaces, recommendation systems, or fraud and bot detection.
You're familiar with e-commerce platforms or large-scale retail and commerce systems, such as Salesforce Commerce Cloud, Shopify, Adobe Commerce, or similar.
You have open-source contributions or a public body of work, for example on GitHub or Google Scholar.

Responsibilities

Shopper Agents: Assist customers and answer engines in finding the products they need across Commerce Cloud storefronts.
Merchant Agents: Help retailers automatically run their storefronts, from merchandising to promotions.
Smart Infrastructure: Automated bot detection and order management that help the enterprise run efficiently at Commerce Cloud scale.
Design and build agentic ML systems that learn from ongoing interaction with users and environments, not just static datasets.
Run applied research experiments end-to-end, from problem framing and data analysis to model design, training, and evaluation.
Prototype quickly and iterate: build hands-on proofs of concept, then help harden the most promising ones for production on Commerce Cloud.
Collaborate with Commerce Cloud engineering, product, and other researchers to translate research results into shipped features.
Stay current with the latest developments in ML and AI research and evaluate their applicability to commerce problems.
Communicate findings clearly through internal writeups, demos, and, where relevant, external publications.