· bigtechsalary Editorial · Career · 5 min read
Adept Ai Research Scientist Salary (2026)
Adept AI Research Scientist compensation in 2026: base, equity structure post-acqui-hire, and how offers compare across agentic-AI labs.
Adept AI Research Scientist Compensation: 2026 Update
Adept’s compensation landscape shifted meaningfully after its 2024 acqui-hire arrangement with Amazon, which pulled a significant portion of Adept’s founding research team into Amazon’s AGI organization while leaving a restructured Adept entity continuing independent agentic-AI research and licensing work. For candidates evaluating a 2026 Research Scientist offer at Adept specifically (as distinct from the Amazon AGI team that absorbed much of the original team), understanding this bifurcated structure is essential to interpreting comp data correctly — much of the publicly cited “Adept” comp data from 2024-2025 actually reflects Amazon’s internal AGI bands, not the standalone Adept entity’s current offers.
This article focuses on verified 2026 offer data for Research Scientist roles at the continuing Adept entity, cross-checked against recruiter disclosures and candidate reports from the current hiring cycle.
The Restructured Comp Model
Post-restructuring, Adept operates as a smaller, more focused research organization with a licensing-revenue business model (its agentic workflow technology is licensed rather than sold as a direct consumer product). This changes the equity risk calculus significantly from the pre-2024 version of the company: equity grants now carry higher uncertainty given the smaller headcount and narrower path to a traditional exit, and candidates report Adept compensating for this with meaningfully higher cash components relative to what the original, VC-flush Adept offered in 2022-2023.
Compensation Bands by Level (July 2026)
| Level | Title | Base Salary (USD) | Signing Bonus | Equity (4yr value, est.) | Total Comp Year 1 |
|---|---|---|---|---|---|
| RS1 | Research Scientist I | $170,000 - $195,000 | $20,000 - $35,000 | $140,000 - $220,000 | $230,000 - $280,000 |
| RS2 | Research Scientist II | $195,000 - $230,000 | $35,000 - $55,000 | $260,000 - $400,000 | $290,000 - $360,000 |
| RS3 | Senior Research Scientist | $225,000 - $265,000 | $50,000 - $80,000 | $440,000 - $650,000 | $380,000 - $470,000 |
| RS4 | Staff Research Scientist | $260,000 - $305,000 | $75,000 - $110,000 | $700,000 - $1,000,000 | $480,000 - $610,000 |
| RS5 | Principal Research Scientist | $295,000 - $345,000 | $100,000+ | $1,100,000+ | $610,000 - $780,000+ |
These bands reflect a cash-heavier structure than comparable roles at Cohere or Mistral — Adept’s base salary ceiling at RS3/RS4 runs 8-12% above Cohere’s equivalent levels, a direct reflection of the higher-risk equity narrative post-restructuring.
What the Research Scientist Loop Evaluates
Adept’s 2026 loop for Research Scientist candidates centers heavily on agentic AI and tool-use research — expect deep technical discussion of action-space modeling, multi-step planning architectures, and evaluation methodology for autonomous agents, alongside a more traditional ML fundamentals round and a research presentation where candidates defend a prior paper or project. Unlike pure LLM-pretraining-focused labs, Adept places unusually heavy weight on RL and decision-making theory given its agentic-workflow focus, and candidates without this background (even strong LLM researchers) sometimes level lower than expected.
A distinguishing 2026 loop feature: candidates are frequently asked to critique a failure mode in an existing agentic system (a screen-recording or trace of an agent failing a multi-step task) and propose a fix — testing applied research judgment over pure theoretical depth.
How Adept Compares to Peer Agentic-AI and Foundation Model Labs
| Company | Comparable Level | Total Comp Year 1 | Equity Risk |
|---|---|---|---|
| Adept | RS3 (Senior) | $380,000 - $470,000 | High, cash-weighted to compensate |
| Cohere | L5 (Senior) | $360,000 - $450,000 | High |
| Mistral AI | Senior Research Eng | $340,000 - $430,000 | High, EU-based tax considerations |
| Anthropic | L5 (Senior RS) | $550,000 - $800,000 | Moderate |
| Amazon AGI | L6 equivalent | $420,000 - $600,000 | Low (public parent) |
The comparison to Amazon AGI is particularly relevant for candidates specifically evaluating Adept versus the absorbed team, since it represents the most direct “what if I’d joined Amazon instead” counterfactual — Amazon AGI’s total comp runs meaningfully higher due to RSU liquidity, but candidates report Adept’s smaller-team environment and direct founder access as a differentiator some are willing to trade cash for.
Negotiation Considerations Unique to Adept’s Situation
- Ask directly about the restructuring’s effect on your specific offer. Because Adept’s 2026 organization is smaller and more focused than its 2022-2023 form, candidates should explicitly ask what percentage of the original team remains and what the realistic equity outcome scenarios look like — vague answers here are a signal to negotiate cash harder.
- Use the Amazon AGI comp comparison directly. Since it’s public knowledge that much of Adept’s original team moved to Amazon, citing Amazon AGI’s total comp bands as a benchmark during negotiation is a legitimate and commonly used lever — recruiters are aware candidates know this and are more prepared to discuss it than a typical “competing offer” conversation.
- Push for higher signing bonus over equity. Given the elevated equity risk profile, converting some of the equity ask into an additional signing bonus or a guaranteed first-year bonus is a more reliably granted request here than at more stable-valuation labs.
For a full framework on how to raise sensitive questions about a company’s stability or restructuring risk without souring an offer, plus scripts for converting equity asks into guaranteed cash, The Big Tech Salary Negotiation Playbook covers this exact scenario: available on Amazon.
Should You Take an Adept Offer Over a Bigger Lab Offer?
The honest answer depends heavily on risk tolerance and career stage. Early-career researchers optimizing for pure compensation and equity liquidity should generally prefer Anthropic, Google DeepMind, or Amazon AGI if those offers are available. Adept makes the most sense for researchers specifically motivated by agentic-AI research depth, direct access to a small, senior team, and comfort with elevated equity uncertainty in exchange for a stronger cash package and potentially faster path to research ownership than a larger organization would offer.
Frequently Asked Questions
Is Adept still an independent company in 2026, or part of Amazon? Adept continues to operate as an independent entity in 2026 following the 2024 licensing/acqui-hire arrangement with Amazon, though a substantial portion of its original founding research team now works within Amazon’s AGI organization. Candidates should clarify which entity a specific job posting belongs to before assuming comp data applies.
Does Adept offer standard 4-year equity vesting? Yes, Adept’s post-restructuring equity grants follow a standard 4-year vest with a 1-year cliff, consistent with typical venture-backed startup structure, though the underlying equity value carries higher uncertainty than pre-restructuring grants issued in 2022-2023.
How does Adept’s RL/agentic focus affect leveling for LLM-only researchers? Researchers with strong LLM pretraining or fine-tuning backgrounds but limited RL/agentic-systems experience are frequently leveled one notch below what their raw seniority would suggest, since the loop weights decision-making and multi-step planning research heavily. Building even a small agentic side project before interviewing has been reported to meaningfully improve leveling outcomes.