At Safeguards, we help financial institutions protect their LLMs against costly risks, starting with hallucinations. Our AI tool can correct inaccurate responses dynamically to build trustworthy RAG solutions for mission-critical use cases. We’re building the corrective layer of Generative AI, directly building solutions to tackle some of the most challenging reliability problems with language models.

Safeguards is well funded by investors like Neo, Forum, and other top VCs in SF who are investors in Scale AI, Mosaic ML, Dropbox, Devin etc. Safeguards’ founder, Kevin Wu, was inspired to start Safeguards after seeing firsthand the challenges of deploying language models and earlier versions of GPT as a Lead Data Scientist at Boston Consulting Group (BCG) for F500s and FI firms. Kevin holds a Master of Science (MSc) after four years of studies at Oxford and attended Harvard as a postgrad, doing AI research at MGH.

Why we’re excited about this space:

We’re solving the “last-mile precision problem” that regulated firms face when deploying mission-critical use cases requiring high accuracy. Given the high cost of non-compliance and financial risk it’s not a matter of if, but a matter of when for every regulated firm to employ corrective safeguards. We’re serving a core business function that is foundational to every firm’s AI strategy and implementation. We believe we have the right solution to serve these teams better to accelerate time to value and mitigate costly risks. Lastly, we get excited about building tools to solve for the most complex and highest-impact challenges, like hallucinations, that underpin the adoption of LLMs.

Who You Are:

As the first Founding AI Engineer of Safeguards, you will partner with the CEO and influence the foundation and growth of our business.

• Proven track record of execution: You have 5+ years of engineering experience, including several years building world-class AI models
• You have hands-on production experience with proprietary LLMs and fine-tuning open-source LLMs with HuggingFace, PEFT, RLHF/DPO, etc.
• Experience with evaluations and running hypothesis-driven experiments to benchmark the performance of LLM and RAG systems
• Care deeply about your craft: you take pride in staying up-to-date with advancements in LLMs and are excited about experimenting with new techniques
• Thoughtful problem solver: you approach problems holistically, starting with a clear and accurate understanding of the context. You think critically about second-order consequences and implications of what you’re building
• Empathetic communicator: communicates nuanced ideas clearly, whether explaining technical decisions in writing or brainstorming in real time. In disagreements, you engage thoughtfully with other perspectives and compromise when needed
• Publication record in ML conferences (ICML, NeurIPS, ICLR, CVPR) is a plus

What We Offer:

• Unlimited growth opportunities to learn and grow into your own role, working on some of the most technically challenging problems of Generative AI
• A driver's seat to building with a world-class startup founder backed by top VCs
• Competitive compensation package (salary, equity, benefits) reflective of a founding member's contribution
• The equipment and resources you'll need to be set up for success working remotely.
• Tuition reimbursement towards AI courses

Safeguards recruits, employs, trains, compensates and promotes regardless of race, religion, color, national origin, gender, disability, age, veteran status, and other protected statuses as required by applicable law.

Safeguards’s dedication to promoting diversity, multiculturalism, equity and inclusion is more than a commitment- it is the foundation of what we do. And thus we strongly encourage people from ALL backgrounds to apply for this position.

Applicants must be authorized to work for ANY employer in the U.S. We are unable to sponsor or take over sponsorship of an Employment Visa at this time.