Member of Technical Staff

Beacon Software•San Francisco, CA

15h

About The Position

Beacon is acquiring and operating a portfolio of vertical SaaS companies. Most private equity firms scale by adding people. We are building Beacon to scale by adding software. The thesis is simple: portfolio operations, value creation, and deal sourcing are bottlenecked by human attention, and an agentic operating system can lift that ceiling by an order of magnitude. We are building that system. A cross-portfolio data lake on open table formats, with a feature store on top that makes the data agent-readable. An action layer that runs workflows across three domains: how we run the portfolio, how we grow the portfolio, and how we acquire into it. A feedback loop underneath that captures every action and outcome with stable identifiers. By the next phase of buildout we will have 100+ portfolio companies running on this platform. That is a problem set with serious data scale, real multi-tenant isolation requirements, and very few precedents to copy from. About the Role Members of Technical Staff (MTS) are the senior engineers who build the platform that everything else at Beacon runs on. You will own a piece of the core stack end-to-end: design, implementation, operations, and the long-term technical direction of that area. This is a Staff Engineer role in everything but name. We run flat. The work is systems engineering at its core. Multi-tenant data infrastructure across very different portcos. Event-driven pipelines that have to be correct under partial failure. Service architectures that have to stay simple as the product surface grows. APIs and SDKs that other engineers — including FDEs out in the field — will build on every day. ML and agentic systems are part of the stack. They sit on top of a foundation that has to be solid first. This is not infrastructure for its own sake. The platform has to be solid before anything else at Beacon works. That is the job.

Requirements

Senior engineering depth (Staff or principal-equivalent).
Experience building and operating systems that real businesses depend on.
Ability to write clean, idiomatic code in at least one of Python, Go, Rust, or TypeScript, and ability to work in any of them.
Experience with distributed systems intuition.
Experience debugging production incidents related to distributed systems.
Experience with idempotency, partial failure, retry semantics, eventual consistency, schema evolution, multi-tenant isolation.
Experience building or operating something non-trivial on a modern data stack (Kafka, Spark, dbt, Iceberg, Snowflake, Databricks, BigQuery, or comparable).
Understanding the difference between a warehouse and a lake, and when each is the right answer.
Platform mindset: building for other engineers.
Comfortable with ambiguity.
Interest in modern ML (not necessarily expertise).
Ability to read a paper, build infrastructure around a model, and have an informed opinion on ML's place in the stack.

Nice To Haves

Prior Staff or Principal Engineer experience at a high-bar engineering org.
Experience with Iceberg, Polaris, Snowflake, or Databricks at scale.
Multi-tenant SaaS or platform infrastructure background.
Production experience with LLM-driven systems, including evals and observability.
Background in offline RL, contextual bandits, or sequential decision-making.
Open-source contributions to data infrastructure, observability, or developer tooling projects.
Shipped LLM-driven systems in production.

Responsibilities

Own one of these areas end-to-end: Data platform, Core services and APIs, Multi-tenant isolation, Workflow and action runtime, Observability and evals, Safety and blast radius.
Design, implement, and operate a piece of the core stack.
Define the long-term technical direction for your area.
Build systems that real businesses depend on.
Build for the engineer two seats over as much as for the end user.
Write clean, idiomatic code in at least one of Python, Go, Rust, or TypeScript.
Structure services and defend design choices.
Debug distributed systems issues, including idempotency, partial failure, retry semantics, eventual consistency, schema evolution, and multi-tenant isolation.
Build or operate non-trivial systems on a modern data stack.
Design APIs that are easy to use correctly and hard to use incorrectly.
Write documentation.
Make migration paths obvious.
Treat developer experience as a feature.
Make decisions that constrain future possibilities.
Build the infrastructure around a model someone else trained.
Have an informed opinion on where ML belongs in the stack and where it does not.