Company Description:
We are creating the future of brain-computer interfaces: building devices now that have the potential to help people with paralysis regain mobility and independence and invent new technologies that could expand our abilities, our community, and our world.
About the Brain-Computer Interface (BCI) Team:
The BCI team is responsible for delivering a product that gives people with paralysis the ability to control computers and phones with their minds at the same speed and functionality level as able-bodied people. In this role you’ll work with neuroscientists, physicians, and software and electrical engineers to develop decoding algorithms that bring the Neuralink device to market.
We are hiring an experienced Machine Learning Engineer to train neural decoders to increase control speed and accuracy, improve reliability, and expand functionality. The machine learning solutions you develop are critical to bring the N1 device to market.
Machine Learning @ Neuralink:
Our ongoing Precise Robotically Implanted Brain-Computer Interface (PRIME) study participants have neural implants with the highest channel count ever put in a human brain. We operationalize research sessions with participants up to seven days a week.
The team has the infrastructure, software engineering resources, and clinical frameworks in place to rapidly move new approaches through benchtop simulations into research sessions. You will be able to rapidly assess neural decoding strategies - often on the same day you wrote the code. If your approach is successful, participants can immediately use your solutions to pursue their life goals.
Neuralink designs all hardware in-house. From custom ASICs to thin-film arrays, there is no part of the technical design that cannot change. Learnings from your work will translate into next-generation device architecture.
Technical Challenges:
If you were to join, here is a representative first project:
Problem: allow a user to perform text entry at 35wpm from the motor cortex
Why does it matter?
- Allows our users to go back to school and get a job.
- Allows users with ALS to communicate 5-10x faster with loved ones.
- Brings hope to hundreds of thousands of folks with SCI + ALS worldwide who have been waiting for a technology like this.
Example Approaches
- Decoding handwriting intention with integrated large language models.
- Designing an alphabet that balances neural separability and intuitive writing.
- Multi-finger typing.
- Innovative software keyboard designs built for BCI.
Challenges:
- Neural Data Instability
- The instability of neural data over time poses a critical challenge for BCI models, often requiring frequent recalibration. The causes of this neural instability remain unclear.
- Data Hungry
- Current approaches require data for each character you want to classify. Ideal methods would scale to the combinatorial space of keypresses that allow efficient computer navigation and text entry without becoming prohibitively expensive to recalibrate/maintain.
- Neural Context
- In our work exploring brain to text, we’ve observed 10% differences in character error rate when the user can see the characters output on the screen vs when the characters are decoded in the background. A 5-10% difference in character error rate translates to 20-40% word error rate.
- User Experience
- Our current modeling approach is sensitive to speed of writing. Deviations from the speed distribution of calibration cause large drops in accuracy.
To learn more about the ML challenges we are tackling @ Neuralink, check out this interview:
Job Responsibilities:
- Understand business and company goals to develop a roadmap and strategies to achieve these goals and meet user needs
- Deliver end-to-end machine learning solutions to improve BCI
- Lead the team by performing at a high standard and setting the bar for how we build and operate our systems
- Ability and desire to mentor others in achieving engineering excellence and to be a catalyst on the team
- Work with cross-functional teams to design new BCI functionalities and novel computer user interfaces
Required Qualifications:
- Evidence of delivering high-impact projects to users or businesses with clear metrics
- Prior experience designing, building, and shipping Machine Learning products for end users
- Deep understanding of machine learning concepts
- Experience in analyzing complex datasets, driving insights, and communicating results in a simple and clear way to both technical and non-technical stakeholders
- Excellent communication and collaboration skills
- Strong coding skills, with a focus on clean, efficient, and scalable code development
Preferred Qualifications:
- Master's or Ph.D. degree in Computer Science, Electrical Engineering, Data Science, or a related field
- Signal processing knowledge
Pay Transparency:
Based on California law, the following details are for California individuals only:
What We Offer:
- An opportunity to change the world and work with some of the smartest and most talented experts from different fields
- Growth potential; we rapidly advance team members who have an outsized impact
- Excellent medical, dental, and vision insurance through a PPO plan
- Paid holidays
- Commuter benefits
- Meals provided
- Equity + 401(k) plan *Temporary Employees & Interns excluded
- Parental leave *Temporary Employees & Interns excluded
- Flexible time off *Temporary Employees & Interns excluded
Multiple studies have found that a higher percentage of women and BIPOC candidates won't apply if they don't meet every listed qualification. Neuralink values candidates of all backgrounds. If you find yourself excited by our mission but you don't check every box in the description, we encourage you to apply anyway!
Neuralink provides equal opportunity in all of our employment practices to all qualified employees and applicants without regard to race, color, religion, gender, national origin, age, disability, marital status, military status, genetic information or any other category protected by federal, state and local laws. This policy applies to all aspects of the employment relationship, including recruitment, hiring, compensation, promotion, transfer, disciplinary action, layoff, return from layoff, training and social, and recreational programs. All such employment decisions will be made without unlawfully discriminating on any prohibited basis.
If you need a reasonable accommodation at any point in the interview process, please let us know. Reasonable accommodations are modifications or adjustments to the application or hiring process that would enable you to fully participate in those processes. Examples of reasonable accommodations include but are not limited to:
- Documents in alternate formats or read aloud to you
- Having interviews in an accessible location
- Being accompanied by a service dog
- Having a sign language interpreter present for the interview