Forward was founded in 2013 by four Stanford Ph.D.s, building the industry's first network digital twin: a mathematically accurate model of the production network. It's the foundation for autonomous networking, giving engineers and AI agents the ability to know the impact of every change before it touches production. That founding instinct still defines how we work. We're accurate and evidence-driven, relentless about clarity, and we'd rather be certain than comfortable, building a groundbreaking platform that transforms how teams run and secure networks across every major cloud and vendor environment.
Global leaders like Goldman Sachs, PayPal, S&P Global, IBM, and Dell trust Forward, alongside fast-growing enterprises and government agencies, realizing an average of $14.2 million in annual benefits, according to IDC. Backed by top-tier investors, including A. Capital, Andreessen Horowitz, Goldman Sachs, MSD Partners, Omega Venture Partners, Section 32, and Threshold Ventures, and headquartered in Santa Clara, we're most proud of our team: curious people who'd rather build what doesn't exist than accept how things have always been done.
Forward is looking for an AI Engineer that will be at the forefront of building intelligent features that make enterprise networks smarter, more secure, and easier to manage. You’ll collaborate with a world-class engineering team, experiment with cutting-edge models and techniques, and see your work directly impact mission-critical systems used by Fortune 500 companies.
This is a role for someone who thrives on innovation, fast prototyping, and bold problem-solving. If you’re excited to push the boundaries of what’s possible with AI—and to do it in a collaborative, hacker-minded team—this is the place for you.
The Role
Forward is building the AI layer on top of the industry's most accurate network digital twin: agents that reason over a mathematically precise model of production networks, so engineers can ask hard questions and trust the answers. As an AI Engineer, you'll build and harden those agentic features and the evaluation systems that keep them honest. Your work ships to some of the largest and most complex networks on the planet, where a plausible-but-wrong answer isn't good enough.
This is a role for a builder who is evidence-driven by instinct: you reach for an eval before an opinion, you read the failing trajectory before you theorize, and you sweat the details because you know that in agent engineering the details are the product. If you’re excited to push the boundaries of what’s possible with AI this is the place for you.
What You’ll Do
-
Improve the quality of shipped agents through eval-driven iteration: error analysis on real trajectories, targeted fixes to prompts, context, and tools, and regression coverage that makes the gains stick.
-
Build new agent capabilities, tools, and surfaces on the Forward platform, with an evaluation in place before they ship and a quality number attached after.
-
Practice context engineering in earnest: retrieval, prompt and tool-result structure, and token-budget management against a large, structured network model that doesn't fit in a context window.
-
Engineer prompts, tool descriptions, and instructions that are unambiguous and well-structured. Communicating clearly to a model is the core craft here, not an afterthought.
-
Help define what “good” looks like for agents completing complex tasks end-to-end.
-
Partner with network domain experts to build and validate evals, and bring the new knowledge to the team through your findings.
You might be a good fit if you
- Have roughly 1 - 2 years of experience in the field
- Have personally owned and shipped at least one real feature on top of LLMs (an agent, a retrieval/RAG system, or an LLM-powered workflow), can point to how you measured whether it was any good: evals, error analysis, A/Bs, quality metrics. You think in terms of "how would we know it's better?"
- Enjoy debugging messy, real-world failures and turning them into improvements.
- Are proficient in one or more programming languages (e.g., Python, Java, C++, JavaScript).
- Possess solid software engineering fundamentals in any language: clean, tested code, methodical debugging, and the judgment to work with minimal supervision and impeccable attention to detail.
- Are comfortable getting productive in a large, unfamiliar codebase.
- Are a strong communicator and collaborator, and are able to work with experts in a domain you're still learning.
Nice to Have
- Experience with eval frameworks, LLM-as-judge, or trajectory / error-analysis tooling.
- Built with agent frameworks or raw APIs (LangChain, LlamaIndex, tool-runner, or hand-rolled tool loops) and can critique their tradeoffs.
- Side projects, open-source work, or writing that shows you track this fast-moving field.
- Any networking, infrastructure, or systems background. It speeds your ramp and helps you write correct evals.
NOTE: Experience in the networking domain is NOT a requirement for this role but is a plus.
Why Forward
- Build AI on a unique substrate: a mathematically accurate digital twin of networks run by the largest enterprises.
- Join an evidence-driven engineering culture where "prove it with an eval" is the norm, not the exception.
- Work directly with the domain experts and a tight, high-signal AI team, with room to grow ownership fast.
- Competitive compensation, benefits, and the chance to shape the future of AI in enterprise networking.
The base pay range for this role is between $180,000 and $230,000. Base pay will depend on your skills, qualifications, experience, and location