Senior/Staff Software Engineer - AI Products

Gem • San Francisco, CA

60d • $190,000 - $260,000 • Hybrid

About The Position

Gem is the only AI-first all-in-one recruiting platform. We're building technology that reinvents how companies discover exceptional talent. When a company hires an exceptional engineer, there's a good chance our system found them first. Over 1,000 industry leaders including Anthropic, Reddit, Figma, Zillow, Robinhood, and DoorDash trust Gem's all-in-one platform to fuel their growth. We've raised $148M from Accel, Greylock, ICONIQ, Sapphire, and Meritech.

What makes Gem special? We have clear product-market fit. Our customers love using Gem every day, and we get to hear directly from them in the morning, then build what they need in the afternoon. It's a rare thing to work on a product people genuinely rely on and appreciate.

We work in our San Francisco office three days per week. In-person collaboration is core to how we ship quickly and build great products together. We offer relocation assistance for strong candidates interested in joining us in SF.

We're hiring senior and staff engineers to join our 5-person AI engineering team. Here's what makes this different: at Gem, software engineers don't just build around models—they work directly on them. You'll fine-tune LLMs and embedding models, rethink our search architecture, clean and optimize data flows, and make calls on what systems and infrastructure we use. We’re even considering next-generation AI hardware (should we move to Groq chips? try Cerebras?) – nothing is off the table if it makes our search faster and more accurate.

This is hands-on work from end to end. You'll integrate directly with LLMs and rerankers, experiment with new models as they launch, build evaluation systems to measure what actually matters, and own the entire stack from Snowflake data pipelines to Elasticsearch queries to the UI someone sees in their browser.

Recruiting is an industry ripe for AI transformation – the features you build will directly help companies discover and hire exceptional talent, impacting their success. Unlike generic ML tooling, the work you do here changes how teams scale and how people land meaningful jobs.

This role offers the chance to work on a small, world-class team (ex-Meta, Uber, Dropbox) building real products in a fast-moving, user-focused environment. You’ll have high ownership, lots of say in how we build, and the opportunity to shape both technology and team norms as we scale. If you’re excited by the idea of shipping something impactful, iterating quickly, and seeing your work change how companies hire – we’d love to talk to you.

Requirements

• 5+ years building production software, ideally full-stack
• Experience with TypeScript, React, and PostgreSQL (or similar)
• Proficiency with Python-based tooling for training, evaluating, or tuning LLMs and embeddings
• You've worked with LLMs, embeddings, vector databases, or search systems in production (not just in a tutorial or research project)
• Strong CS fundamentals: data structures, algorithms, databases
• You ship features end-to-end without waiting on others to productionize your work
• You thrive on small teams where everyone touches everything
• You start with the problem, not the solution. You're comfortable with ambiguity and figure out what actually needs to be built.
• You know when to use the fancy new model and when Postgres is good enough. You care about craft but you ship.
• You explain complex technical decisions clearly. You're opinionated but not dogmatic.
• You like moving fast. You want to experiment with a new model in the morning and see it in production by afternoon.
• You care about the product. You think about how recruiters actually use search, not just how to make the algorithm better in isolation.
• You mentor teammates and contribute to raising the bar for technical quality across the team.
• You collaborate closely with product managers and designers to create cohesive end-to-end features and user experiences.

Nice To Haves

• You've built agents or worked on eval infrastructure
• You know your way around vector databases (Pinecone, Weaviate, Qdrant) or search systems (Elasticsearch)
• You've worked with RAG architectures or ML observability
• You've worked at an early-stage startup where you had to figure things out yourself
• Background in information retrieval, ranking, NLP, or recommendation systems
• Experience with data pipelines (Snowflake, dbt), model deployment, or monitoring models in production
• Comfortable in Python and working with ML frameworks like PyTorch or TensorFlow

Responsibilities

• Model work – fine-tuning LLMs and embedding models for recruiting queries, testing new providers as they launch, building systems to evaluate what actually improves search quality
• Search at scale – making semantic search instant across 800M+ profiles, integrating rerankers to surface better candidates, designing the feedback loops that help search get smarter
• Data infrastructure – owning pipelines in Snowflake that feed our models, cleaning and structuring candidate data, building the systems that let us experiment quickly without breaking production
• Shipping full-stack features – writing the code from prompt engineering to UI, creating interfaces that make complex search feel simple, iterating based on what recruiters actually tell us

Benefits

• 10-year stock option exercise window
• Flexible Time Off and 16 paid holidays (including company wellness days)
• Best-in-class medical, dental, and vision coverage
• $1,200 annual learning and development stipend
• 16 weeks paid parental leave for all parents, plus $1,500 new-parent perk and flexible return-to-work options

What This Job Offers

• Job Type: Full-time
• Career Level: Mid Level
• Education Level: No Education Listed