San Diego Machine Learning (SDML) has been bringing together the machine learning community, hosting technical talks, ML book readings, Kaggle competition teams, and ML office hours for seven years. Through participation in SDML, many members have improved their skills and advanced their careers. The SDML leadership team has created a new educational program to condense this process and dramatically speed up the learning process for dedicated individuals.
The SDML Kaggle Internship Program (SDML-KIP) is a program designed to allow data scientists and ML engineers to rapidly improve their skills and demonstrate their knowledge. Participants will work under two mentors for between two and three months, working as a team on a real Kaggle competition. In addition to learning key machine learning concepts, such as cross-validation, participants will get to see these concepts in action and observe how they help the modeling process, such as improving correlation between local validation and the public leaderboard. Many of the principles to be taught during the SDML-KIP program were piloted with Cohort 0 during the Vesuvius Challenge - Ink Detection competition, in which the team came in first place. (We cannot promise that SDML-KIP participants will win every competition!)
The SDML Kaggle Internship Program is for new to mid-level data scientists and ML engineers seeking to gain real-world skills and prove their knowledge and expertise. Participants are expected at a minimum to be able to program in Python and already know the basics of machine learning, such as supervised learning and how to train a logistic regression model. Cohorts will be selected with participants at a similar skill level. Depending on the applicant pool, a given cohort may be chosen at a more junior or a more mid-level experience level.
Participants will be expected to spend a minimum of ten hours per week for the 2-3 month duration of the program. Time will include participation in weekly meetings on Saturday afternoon and work throughout the week on the ML competition. Participants will be required to run and modify code, read and post on Kaggle discussion boards, participate in weekly meetings, and engage with mentors and other participants on Slack. At the end of the program, participants will create a final report, documenting the skills and techniques they used during the Kaggle Competition.
Cohort 1 of the SDML Kaggle Internship Program will begin in early- to mid-May 2024, and run through July 15, 2024. Interns will be working on the LEAP - Atmospheric Physics using AI (ClimSim) competition (https://www.kaggle.com/competitions/leap-atmospheric-physics-ai-climsim). Weekly working sessions will be on Saturdays, starting at 2:15pm. Additional working time with mentors or other participants will be worked out individually. Participants will be expected to do additional work on their own each week. Total time each week for participants is expected to be ten hours per week. Candidates should apply for a future cohort if they are not able to commit to the time commitment for the duration of the program.
Participation in the SDML-KIP program is free. Interns will be unpaid and will not receive any stipend. If the team wins a prize in the Kaggle competition, participants will be able to keep their share of the prize money. Participants will need access to computing resources. There are many free resources online at Kaggle and elsewhere. Some participants may voluntarily choose to purchase cloud computing resources, but it is not expected that anyone would spend more than $50.
Candidates for SDML-KIP cohort 1 should fill out the following interest form as soon as possible: https://docs.google.com/forms/d/e/1FAIpQLSfucPXr4NUBBp_PTx2HzM5SEt37RKbXuuRQqDovqD-UYhpxpg/viewform?usp=sharing. Decisions will be made on a rolling basis, starting Sunday, May 5. Candidates will be interviewed by one or more mentors. To test their experience level, candidates may be asked to explain machine learning concepts or perform some coding tasks. Up to three applicants will be accepted for Cohort 1. Applicants will be notified by May 15, 2024, possibly sooner.
San Diego Machine Learning is proud to have the following two experts as mentors for SDML-KIP Cohort 1:
Ryan Chesler
Ryan is Principal Data Scientist at H2O.ai. He is a Kaggle double grandmaster who has competed in over 100 competitions. He has trained over 10,000 deep learning models and applied them to countless domains. Ryan is self-taught, with much of his experience coming from participating in Kaggle competitions.
Ted Kyi
Ted is VP of Data Science for Deep Sentinel. He is a Kaggle competitions expert. Ted has a degree in computer science and has been participating in Kaggle competitions for seven years. He has extremely strong technical and practical skills, and he is skillful at simplifying and teaching complicated concepts intuitively and rigorously.
We have started to build a FAQ as questions have come in. Please see the separate FAQ document for additional information.