Compute Trader

Beam is building the fastest cloud runtime for AI. Our serverless runtime launches GPU-backed containers in under a second, powering apps that scale to millions of users. Ambitious startups and Fortune 100 companies use Beam to host custom ML models and run LLM-generated code in secure sandboxes. We also provide dedicated GPU clusters to AI companies that need bare-metal capacity. That capacity comes from a network of datacenter partners and, increasingly, from our own hardware.

Beam serves millions of requests a day for hundreds of companies in production. We’re backed by Y Combinator, Tiger Global, and prominent developer-tool founders, including the founder of Snyk and the former CTO of GitHub.

About the role

AI companies that need a few dozen to a few hundred GPUs have a hard time getting them. The largest providers prioritize bigger deals, and the rest of the supply is spread across hundreds of datacenters worldwide. Beam closes that gap.

We’re hiring a Compute Trader to match customer demand with that supply. You’ll find capacity across our own hardware and our partners’ datacenters, negotiate the terms, and make sure we can deliver what we sell. You’ll report directly to our CEO.

What you’ll work on

• Match demand to supply. Take customer requests from our sales team and find the right capacity at a price and term that work for both sides.
• Source and onboard suppliers. Build relationships with neoclouds, independent datacenters, and OEMs around the world. Run diligence, KYC, and export screening before anyone signs.
• Verify before we quote. Confirm GPU type, location, network fabric, operator, and delivery date for every deal. Know how suppliers differ, technically and commercially, and which customers each one fits.
• Structure the deal. Negotiate price, term, deposits, and prepayment with suppliers.
• Keep our fleet full. Maximize utilization of our own clusters, flag renewals early, and re-source capacity before contracts end.
• Read the market. Track pricing by GPU, term, region, and deposit. Turn it into a view on where to buy, when to lock in, and when to own instead of rent.
• Build the process. Create the capacity tracker, pricing book, and supplier scorecards we’ll use as volume grows.

What we’re looking for

• 2+ years in a sales or trading role, GPU cloud sales, capacity sourcing, datacenter operations, or a related field
• You negotiate hard and hold suppliers to what they promised.
• You verify before you promise. You’d rather slow down a quote than sell capacity we can’t deliver.
• You’re comfortable with ambiguity and move fast without dropping details.

Nice to have

• An existing network in the compute market, whether providers, brokers, datacenter operators, or buyers
• You understand what drives GPU pricing. You can explain why two quotes for the same GPU differ by a dollar an hour.
• You can read a spec sheet and a contract. You know why a customer cares about InfiniBand versus RoCE, and how a deposit schedule affects cash flow.

Benefits

• Competitive salary with meaningful equity
• Health, dental, vision, 401(k)
• Join a fast-growing pre-series A company at the ground floor
• Fitness stipend, learning budget, quarterly team offsites