
Machine Learning Research Engineer, NLP, Input Experience
Immigration summary
Visa sponsorship
This employer sponsors this kind of work repeatedly, and is still filing this year.
5,644 recent H-1B filings · 8 similar-role filings
View visa evidenceGreen card sponsorship
This employer has recently sponsored green cards at scale.
1,278 recent certified PERM filings · 112 similar-role filings
View green card evidenceJob description
From our origins in iPhone keyboard input, the Input Experience NLP team has expanded our broad charter: enhancing the user experience with robust language understanding and personalized text composition, across all Apple platforms and languages. Generative AI is a transformative technology, and we are just beginning to harness its potential to help users digest information and express themselves more clearly. On our team, you will help build the future and shape its evolution. Our team is responsible for key Apple Intelligence portfolios such as personalized Writing Tools, Summarization (Mail, Messages, Notifications, etc.) , Found In, Smart Actions as well as the entire keyboard backend: autocorrection, inline completions, proofreading, across all Apple platforms. Building on years of innovation in intelligent systems and on-device machine learning, we are now scaling efforts in bringing powerful foundation models (on-device and server) directly into everyday workflows.
We are looking for an engineer who can work at the intersection of ML, NLP and software engineering, specifically focused on innovating and evolving our data, tooling, modeling and evaluation pipelines with agentic harnesses, to scale globally. We are shifting the entire paradigm of ML product development across feature definition, data synthesis, model training, auto-evaluation, model probing, evaluation and user feedback to agentic workflows. You will have the opportunity to define and execute state-of-the-art paradigm for a swathe of high-impact features and languages, creating ML playbooks for scale, and influencing the rest of Apple. You will also be responsible for building and refining the personalized agent trajectory pipelines for data and evaluation across synthetic personas and languages. The role provides an opportunity to join an ambitious, collaborative team in a unique position to bridge the gap between cutting-edge ML research and features used by millions. You will work closely with cross-functional partners in human interfaces, user studies, internationalization, and system integration. You are not just developing technology; you are crafting experiences that feel like magic to the end user.
Sponsorship evidence
Why Openbound reached the conclusions above.
Visa sponsorship evidence
Current posting
Silent on sponsorship
Employer H-1B history
- 5,644
- recent certified H-1B filings
- 2,440
- new-hire petitions
- 8
- filings for similar roles
- 5,619
- so far in FY2026
Filed titles like this role: machine learning research engineer · ml research engineer
More evidence details
- 8 certified H-1B filings for this same role, 2,440 new-hire petitions across the employer, and 2,837 new hires already this year
- 8 certified H-1B filings for this same role
- 5,644 recent certified H-1B filings across the employer
- Still filing this year — 5,619 filings in FY2026
- 2,048 USCIS H-1B new-employment approvals, counted separately from LCA filings
- 9,550 further USCIS approvals for extensions or transfers
- The posting says nothing about sponsorship either way
- No filing activity is recorded for this job's location
Green card sponsorship evidence
Employer PERM history
- 1,278
- recent certified PERM filings
- 112
- filings for similar roles
Filing history reflects past employer behavior; it isn't a promise for this opening.
All open roles at AppleHow Openbound evaluates sponsorship
Visa history uses official U.S. Department of Labor H-1B LCA disclosure data and USCIS H-1B petition history. Green card history uses DOL PERM disclosure data. Each is read for the employer as a whole, for roles like this one, and for this location, weighted toward the most recent fiscal years.
An employer is matched to its filing entities by verified legal name and reviewed aliases; a match is never made on a name resemblance alone. Where no verified entity can be matched, the page says so and draws no conclusion from the absence. 1,053 filing titles were examined for this employer.
What this posting states outranks history in both directions, and an employer's published policy outranks past filings. Filing history reflects past behavior; it is not a promise of sponsorship for this opening, and none of this is legal advice.