Job Description
The next wave of competitive advantage isn't a better general model — it's a model that knows your business. Eragon builds company-specific AI that you own, trained on your data, deployed in your environment, and that never leaves it. Our models don't just answer questions — they learn from every interaction, getting smarter the longer they're in the field.
We've built a novel reinforcement learning framework — RLQF (Reinforcement Learning from Query Feedback) — that creates a compounding improvement flywheel: every query, correction, and workflow becomes training signal. The result is an AI system that keeps pulling ahead of static fine-tuning or RAG-based approaches over time.
We're backed by $10M in seed funding and already generating revenue with a fast-growing customer base.
The Role
As an Applied Research Engineer at Eragon, you'll train and deploy company-specific models for some of the world's most sophisticated enterprises. You're not tuning general benchmarks — you're building AI that powers real business workflows, learns from real user feedback, and delivers measurable outcomes.
You'll embed directly with customers, understand how their organizations think and operate, and own the end-to-end process: from data ingestion and environment design, through RLQF training cycles, to production deployment.
What You'll Do
Design, train, and deploy customer-specific models using Eragon's RLQF framework
Own customer engagements end-to-end — from data architecture to production rollout
Build reinforcement learning environments, evaluation pipelines, and reward models tailored to each customer's domain
Work directly with technical and business stakeholders to surface feedback signals and close the training loop
Translate what you learn in the field into improvements to Eragon's core platform and training stack
Collaborate across research, product, and infrastructure to define architectures and best practices
What We're Looking For
Meaningful AI/ML research experience, in industry or through substantive academic work
Deep comfort with training and fine-tuning models on real-world, messy enterprise data
Experience with RL — environments, reward modeling, RLHF/RLQF, or evaluation frameworks
Strong problem-solving instincts in ambiguous, fast-moving environments
Ability to work directly with customers: listen well, understand their workflows, and translate needs into training signal
Drive to own problems end-to-end and ship outcomes, not just experiments
Strong Candidates Also Have
Production experience fine-tuning or post-training language models at scale
Background building agentic or tool-use systems
Experience at a forward-deployed or customer-facing engineering organization
Prior founder or early-stage engineer experience
Published work, open source contributions, or side projects that demonstrate technical range
Logistics
Based in San Francisco. Competitive salary and equity, full health benefits, unlimited PTO, daily meals, and relocation support. We sponsor visas.