Asks for

The ad says“Comfortable with LLM-native engineering. You have shipped real software using foundation models — not toy demos. You know how to write evals, when to fine-tune, how to keep an agentic workflow from going off the rails, and where the autonomy tier should sit for any given action class.”
The ad says“Comfortable with LLM-native engineering. You have shipped real software using foundation models — not toy demos. You know how to write evals, when to fine-tune, how to keep an agentic workflow from going off the rails, and where the autonomy tier should sit for any given action class.”
The ad says“Comfortable with LLM-native engineering. You have shipped real software using foundation models — not toy demos. You know how to write evals, when to fine-tune, how to keep an agentic workflow from going off the rails, and where the autonomy tier should sit for any given action class.”
The ad says“Comfortable with LLM-native engineering. You have shipped real software using foundation models — not toy demos. You know how to write evals, when to fine-tune, how to keep an agentic workflow from going off the rails, and where the autonomy tier should sit for any given action class.”
The ad says“A strong generalist engineer. You can ship full-stack TypeScript or Python on the backend, React on the frontend, SQL fluently, comfortable in someone else’s codebase within a day. You do not need to be a specialist in any one stack. You need to be the person who can land the whole thing.”
The ad says“A strong generalist engineer. You can ship full-stack TypeScript or Python on the backend, React on the frontend, SQL fluently, comfortable in someone else’s codebase within a day. You do not need to be a specialist in any one stack. You need to be the person who can land the whole thing.”
The ad says“A strong generalist engineer. You can ship full-stack TypeScript or Python on the backend, React on the frontend, SQL fluently, comfortable in someone else’s codebase within a day. You do not need to be a specialist in any one stack. You need to be the person who can land the whole thing.”
The ad says“A strong generalist engineer. You can ship full-stack TypeScript or Python on the backend, React on the frontend, SQL fluently, comfortable in someone else’s codebase within a day. You do not need to be a specialist in any one stack. You need to be the person who can land the whole thing.”

About Beacon Software

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. It has two pieces. A sensing layer is a cross-portfolio data lake on open table formats with a feature store on top that makes the data agent-readable. An action layer runs workflows across three domains: how we run the portfolio, how we grow the portfolio, and how we acquire into it. Underneath both sits a feedback loop that captures every action and outcome with stable identifiers.

Beacon has raised $550M+ from investors including General Catalyst, Lightspeed, D1 Capital, CPMG, and the family offices of the founders of Stripe, DoorDash, and Ramp.

About the Role

Forward Deployed Engineers (FDEs) are the engineers who go where the work is. You will be embedded with one or two portfolio companies at a time — sitting with the GM, the operators, and the customer success lead — figuring out where the leverage is, and shipping agentic software directly into their operation. The platforms exist. You are the person who turns them into outcomes on a specific business.

This is a Palantir-style FDE role applied to private equity operations. You will spend a meaningful fraction of your time on-site or in the portco’s tools, and the rest of your time back at Beacon HQ pulling the patterns you learned back into the platform. The best FDEs at Beacon will not just deliver for the portcos they are deployed at — they will compress what they learned into reusable playbooks that ship to every other portco in the same vertical.

This is not a consulting role. We ship software, not slides. The deliverable is a running system.

What You'll Do

Embed deeply with portcos. Spend the first few weeks at a new portco understanding the operating model from the inside: what the GM cares about, where revenue actually comes from, where the team’s time goes, where the system is leaking. Earn the trust of the operators before you start changing things.
Find the leverage. Identify the 2–3 workflows where agentic software will meaningfully change the trajectory of the business — churn-save, pricing experiments, lead enrichment, support deflection, AR follow-up, bug-fix PRs, weekly pulse, whatever the specific portco needs. Pick the wedge, not the wish list.
Ship the wedge. Build the integrations, the prompts, the evals, the workflows, and the UI surfaces required to land the first agentic loop at that portco. Use the platforms our core engineering team is building, and extend them when they fall short.
Operate the loop. Stay close to the system after it ships. Watch the outcomes. Tune the prompts. Adjust the autonomy tier as the system earns trust. Hand off cleanly to portfolio ops once the loop is stable.
Compound the learnings. Bring the patterns back to HQ. Codify what worked into playbooks that ship to every other portco in the vertical. The 11th deployment in a home-services SaaS should be 10x faster than the 1st, and that compression is your job as much as the engineers building the platform.
Carry the GM relationship. You are the technical face of Beacon to the portco GM. Build the trust, set the expectations, and make sure the GM sees the system as their leverage — not Beacon’s audit tool.

Who You Are

A strong generalist engineer. You can ship full-stack TypeScript or Python on the backend, React on the frontend, SQL fluently, comfortable in someone else’s codebase within a day. You do not need to be a specialist in any one stack. You need to be the person who can land the whole thing.
Customer-obsessed. You like being in the room with the operator, the GM, the customer success lead. You take the problem they actually have, not the problem they described. You are not interested in building platforms that no one uses.
High autonomy, low ego. You will be the only Beacon engineer on the ground at a portco most of the time. You need to make calls without a committee, defend them when challenged, and change your mind when you are wrong.
Comfortable with LLM-native engineering. You have shipped real software using foundation models — not toy demos. You know how to write evals, when to fine-tune, how to keep an agentic workflow from going off the rails, and where the autonomy tier should sit for any given action class.
Operator instincts. You can read a P&L. You understand why a churn-save matters more than a feature launch this quarter. You know that a 5-point improvement in support deflection is worth more than a 50-point improvement in code-review velocity, depending on the business.
Strong written communicator. Beacon runs on writing. You will write the wedge memo, the GM update, the post-deployment retro, the playbook that ships to every other portco. If your writing is fuzzy, the system stays fuzzy.

Bonus Points

Prior FDE experience
PE, consulting, or M&A diligence exposure, particularly the post-LOI integration cadence
Comfort with Iceberg-based data platforms (Snowflake, Databricks) and event-driven systems
Built or operated agentic systems in production, including evals, autonomy gating, and outcome tracking

Our Values at Beacon Software

Humility: We acknowledge that the path to getting to the right answer involves being wrong along the way. We have strong beliefs which are weakly held. We actively seek new ideas and believe we can learn from anyone at any time.
Honesty: We are truth seeking in our approach to business problems. Business is a repeat game and we believe that human relationships generate alpha. We understand that trust is earned over a lifetime and can be lost in an instant.
Hunger: We play to win. We hold ourselves to high standards and will not be outworked. We take pride in having a deep sense of responsibility to ourselves, each other, our partners, and our customers. We believe to whom much is given much is expected.
Horizon: We seek to build a generational software company. This will take decades. We manage our expectations and those of our partners to take advantage of the 8th wonder of the world - compounding growth.

How We Use AI in Our Hiring Process: To ensure transparency, we want candidates to know that Beacon Software uses Artificial Intelligence and AI-enabled tools to assist with screening, reviewing, organizing and highlighting profiles and applications that match the key requirements for each role.

AI does not make hiring decisions: Every application is reviewed by a member of our team, and all decisions throughout the process are made by humans. We use AI to support efficiency and consistency, not to replace human judgment. We are committed to a fair, thoughtful, and equitable experience for every candidate.