# AI Researcher at Varick Agents > Location unclear · Unspecified · Software engineering · open. Source: https://alderhq.app/jobs/ashby%3Avarick-agents%3A1a6e4e7d-f1b2-471e-bf29-dc6da60bbfbd (read 2026-11-07). ## Facts (observed) - Company: [Varick Agents](https://alderhq.app/companies/varick-agents) - Location text: SF - Where it can be done from (inferred): Location unclear (evidence: "Location: SF") - Posted: 2026-07-29 (102 days ago) - First seen by alder: 2026-11-02 · last verified: 2026-11-03 - Reposted: 0 times · edited: 0 times - Apply on the employer's site: https://jobs.ashbyhq.com/varick-agents/1a6e4e7d-f1b2-471e-bf29-dc6da60bbfbd - Skills listed (inferred): AWS, LLMs ## Pay and sponsorship - Posted pay: $200k–$250k per year (from the listing's pay field) - Market band for similar posted roles (SF Bay Area, n 1752): p25 $161k, median $200k, p75 $239k ## Description (the employer's text) ### AI Researcher @ Varick Agents ### ABOUT VARICK Varick builds AI agents that take over real operational workflows inside the world's largest enterprises. Our forward-deployed engineers and strategists embed with client teams to map how work gets done and build POCs, then our engineers bring those agents into production. We're venture-backed, revenue-generating from day one, and already in production inside several of these companies. You'll join an elite team from Meta, AWS Bedrock, Citadel Securities, McKinsey, BCG, Stanford, and more. ### ABOUT THE ROLE You make our agents better at the actual work. You own the post-training loop: taking real workflow data from our deployments and turning it into models and evals that push agent accuracy where a demo-grade model fails on the tenth run. This is applied, production-facing research. Your results ship to agents running inside the world's largest enterprises. ### WHAT YOU'LL DO Own fine-tuning, preference optimization, and post-training pipelines that lift agent performance on real enterprise workflows Build the eval suites and golden sets that measure whether a change actually helped, before it ships to a client Turn traces and failure cases from production into training data and targeted improvements Work with the platform team on the model layer: routing, distillation, and cost and latency budgets Keep a clear view of where fine-tuning earns its keep versus better prompting, retrieval, or scaffolding ### WHAT WE'RE LOOKING FOR Strong applied ML background with hands-on post-training experience (SFT, RLHF, DPO, or similar) on LLMs At least one model or system you took to real users, and a clear account of how it behaved in production Fluency with the modern training stack and a working grasp of inference economics Evaluation discipline. You trust measured results over vibes ### HELPFUL EXPERIENCE Published research or open-source work in LLMs and agents Agentic evals, tool-use training, or long-context work Reinforcement learning environments Taking research into production ### CULTURE Extreme ownership. You own the tasks end-to-end, at an extreme level. You think beyond the task at hand, consider upstream and downstream impact, and are accountable for the entire outcome An instinctual sense of urgency. You thrive in and help foster an environment with urgency. When a new request comes in, it is instinctually clear to you how that fits into the existing prioritization list. If it goes to the top of the list, you jump on it immediately, help the rest of the team realize the shift, and consider any impact of de-prioritizing other items Low ego, high agency. We are all on this rocket ship together. If you win, the team wins and vice versa. You don't wait for permission to pursue a task. If you can do something that moves the needle, do it Curious. Our work requires deep curiosity, both when it comes to understanding the clients' work and what matters to them, and when it comes to keeping up with and adopting the latest AI developments. Examples to showcase this could be pursuing niche interests, reading textbooks on the weekend, questioning the latest research, etc. Perfectionist. You must be proud of your work product. In a client-facing business, it is critical that we put correct, well-informed, and well-documented material in front of the client. We are held to the highest standards and embody that mindset internally within Varick too Startup mindset. Things are ambiguous. The hours are long. Weekend work is expected. And this excites you. With AI, the world is moving extremely quickly, and the ways of working are changing every day. We are at the forefront and help facilitate this rapid change. In the next 3, 5, and 10 years, things are going to change significantly, and you join to work hard and be a part of this exciting journey towards an AI-native future Feedback metabolism. You never need the same correction twice, and you treat every review as compounding rather than criticism. You spot opportunities to give, and are comfortable giving, feedback to all of your peers, be it upward, downward, or horizontal --- Data: alder (https://alderhq.app). Cite https://alderhq.app/jobs/ashby%3Avarick-agents%3A1a6e4e7d-f1b2-471e-bf29-dc6da60bbfbd and the date you read it. Observed facts come from the employer's own listing; location reach, level, family and skills are inferences.