NLP Researcher (B.S./M.S./Ph.D.) (Minneapolis, MN; Boston, MA)
NLP Researcher
Location: Minneapolis/Boston
Note: Due to the nature of our contract work, positions at SIFT are restricted to US citizens and permanent residents. Also, this position requires that you be fully vaccinated against COVID-19, unless you require a reasonable accommodation for religion or a health-related need.
SIFT cooperates with universities and companies around the world on cutting-edge NLP technology. Our diverse-- and growing-- set of NLP projects include:
- Analyzing academic articles for reproducibility.
- Extracting information from scientific articles to enhance computational models of cancer.
- Analyzing social media for hate speech and dehumanizing language.
- Building dialogue systems with biologists and analysts to simulate and analyze data.
- Assessing implicit linguistic bias to predict and characterize social inequalities.
We're seeking bright, passionate individuals to join us in these endeavors and improve the state-of-the-art in theoretical and applied NLP!
What will you do as an NLP Researcher at SIFT?
As an NLP Researcher at SIFT, you will have the opportunity to:
- Prototype state-of-the-art NLP models for semantic parsing, information extraction, classification, and sentiment analysis.
- Use technologies including pytorch, transformer architectures, spaCy, and SIFT's state-of-the-art NLP libraries.
- Utilize psychosocial and linguistic theories from social science to design and implement NLP models.
- Integrate NLP outputs with large knowledge graphs to help build multi-university, multi-disciplinary AI systems.
What kind of background is needed for this role?
Basic qualifications to be an NLP Research at SIFT include:
- Background and formal training in NLP including computational linguistics and/or machine learning.
- Proficiency with Python programming and statistics libraries.
Preferred qualifications include:
- Experience with information extraction including named entity recognition (NER), relation extraction, and knowledge graphs.
- Familiarity with transformer-based language models, including pre-trained autoregressive and/or masked models such as BERT, BART, T5, ROBERTA, GPT-2/3, etc.
- Familiarity with collaboration tools such as Jupyter Notebooks, Matplotlib, TensorBoard, etc.
- Good practices with sharing code easily with larger teams using docker containerization.
- Experience with building datasets, working with annotators, measuring annotator agreement, writing annotator guidelines.
What are some of the benefits of working at SIFT?
- Liberal Flex Time Policy: you can work in a way that makes sense for your work/life balance.
- Salary is open to negotiation and will be competitive.
- Health, Dental, Life, and Disability Insurance (SIFT pays 80% of health insurance premium and 100% of life insurance premium).
- 20 paid vacation days per year, plus unlimited sick days.
- 401K matching funds (5% of salary, vested immediately; options of traditional 401k, Roth 401k, or combination, with access to 50+ mutual funds).
- Bonus program for technical merit/contribution.
- Computer of choice (Mac, Windows, or Linux) provided by SIFT.
- Home Office - $2,000 allowance/year.
- Full internet and cell phone reimbursement.
- Profit sharing (as 401k distributions or additional vacation days).
- Tuition reimbursement for continuing education (50% of tuition paid, up to $10,000 per year).
- Parking and bus card reimbursement.
Working at SIFT
SIFT is a unique company that values innovation and employee empowerment. As a small, but stable, business SIFT provides an efficient and thin organizational structure that helps employees focus upon doing interesting technical work without significant red-tape.
SIFT's cooperative-style work environment provides employees a platform to set and pursue their professional interests, including opportunities to focus upon engineering, basic research, project management, business development, IT, and commercialization.
You can apply through this Handshake post, or simply email your resume to hiring@sift.net!