AI Researcher / Engineer / Intern at FAME INC

New York, New York, United States -

Full Time

Start Date

Immediate

Expiry Date

22 Aug, 26

Salary

300000.0

Posted On

24 May, 26

Experience

0 year(s) or above

Remote Job

Yes

Telecommute

Yes

Sponsor Visa

Yes

Skills

Machine Learning, Multimodal Data Modeling, Self-supervised Pretraining, AI Coding Tools, Experiment Tracking, Dataset Versioning, Evaluation Protocols, Model Architecture Design, Python, Signal Processing, Research Tooling, Rapid Prototyping

Industry

technology;Information and Internet

Description

Hi, I'm Brian, co-founder of Egra. This posting is one role with wide possibilities: full-time researchers, full-time engineers, interns, college kids who've been shipping since they were 16. The bar is the same. You'll have complete ownership over your work from day one. No lengthy onboarding, waiting for permission, or navigating layers of approval. Real problems, real compute, real autonomy. The work you do in your first month will be in the product. What you'd be doing We're training models on the highest-bandwidth data anyone has tried to model at scale. Biosignals, eye-tracking, pupillometry, video of human faces, audio, the content humans are reacting to, downstream behavior. A non-exhaustive list of projects on the table: Designing self-supervised pretraining objectives on multimodal physiological + content data Stress-testing recent multimodal / signal foundation model papers to understand exactly where and why they break under distribution shift Building evaluation protocols that distinguish real progress from leaky-benchmark noise Shipping the internal research tooling (experiment tracking, dataset versioning, agentic eval pipelines) that lets research compound instead of repeat Closing the loop between offline model results and the live product Writing the internal research memos that become the shared knowledge base: "why model X fails on dataset Y," "what we tried and why it didn't work" You should be able to take any one of these and have something running end-to-end in a week. We don't separate "research" and "engineering". The people we want are both. Where this is going We're building toward a future where AI is actually optimized for what it does to the human on the other side of the screen. The current signal AI is trained on is a shadow of what matters. Research culture A few strong opinions: Bitter Lesson by default. We're skeptical of hand-engineered features and domain-specific heuristics. We'd rather let models discover structure than force-feed it. Reproducibility over vibes. If we can't tell you which preprocessing version produced a result, we don't trust the result. Failed experiments are documentation, not waste. Write up what doesn't work with the same care as what does. AI-augmented everything. Claude Code, Cursor, internal eval agents, generated tooling. Anything one person can do with AI tools is owned by one person. Who we're looking for The shape we want, in priority order: You can take a vague research direction and ship something concrete within a week. You have strong opinions about what makes representations actually generalize, and they're informed by experiments you ran yourself. You're comfortable with heterogeneous, multimodal data and you have a toolkit for making it useful. You ship fast with AI coding tools. Codex, Claude Code, agents. You have taste. You can look at a benchmark and tell us why it's leaky. You can look at a model architecture and tell us what it can't possibly learn. Background doesn't matter much. PhD, no PhD, dropped out, finishing undergrad — all fine. We've hired interns who out-shipped postdocs and we'll do it again. You should NOT apply if: Your value proposition is "I know EEG" or "I have a neuroscience background." This is an ML role, not a neuro role. We are not hiring for domain expertise in any signal modality. We're hiring for general modeling ability on hard data. You need a clear roadmap or a manager to do your best work. You're more interested in neuroscience theory than building systems that work in production. You rely heavily on hand-crafted features, classical signal-processing pipelines, or domain-specific engineering as your main contribution. You want to publish first and ship second. Interview process Three conversations, total ~90 minutes: 30-min intro. We tell you what we're building, you tell us what you've built. Casual, no prep. 30-min technical jam. We want to see how you think about problems, approach them, and work with us on a day-to-day basis. 30-min deep dive with both founders. Past work, taste, whether we'd enjoy working together every day. Benefits Competitive salary and meaningful equity (full-time); top-of-market intern stipend Platinum-tier health insurance Uncapped compute access. Uncapped AI tooling budget. Full research autonomy No bureaucracy, no review committees Relocation and visa support; in-person strongly preferred, flexible on edge cases

Responsibilities

Develop and train multimodal models using high-bandwidth physiological and content data. Build internal research tooling, evaluation protocols, and integrate offline model results into the live product.