Cog*Works is a pre-college course offered by the MIT Beaver Works Summer Institute aimed at teaching the fundamentals of machine learning and artificial intelligence. Topics include fundamentals of machine learning, audio processing, computer vision, and natural language processing. The course includes weekly capstone projects with a final project and showcase at the end of the course.- Developed and delivered lectures to a class of 32 high schoolers at an MIT pre-college program in machine learning and AI.- Worked with and supported small groups of students in developing capstone projects related to the course.

Problem: Underrepresented middle and high schools in the Milwaukee area wish to provide STEM opportunities to their students by participating in FIRST Robotics but need additional support in running them. FIRST Ambassadors fill the gap by mentoring those robotics teams, bringing their technical experience at the college and professional level directly to schools in need.- Mentored underrepresented middle and high schools in the Milwaukee area participating in any of the FIRST Robotics programs.- Provided 10 years of experience in matters regarding embedded systems programming, paper writing, and presentation.- Prepared students for their future academic and professional careers beyond high school.- Currently mentoring one of my former FIRST Robotics Competition (FRC) teams: MORE Robotics, with ~15 students per year.

Problem: Students, especially new ones, require additional support to find success in their classes for a variety of reasons. RCAS Tutors fill the gap by providing weekly 1:1 and group tutoring sessions for specific classes with the intent of developing students into active learners who can succeed outside of the tutoring session.- Tutored ~75 students over 15 sessions per year in mathematics, computer science, and technical communication classes.- Prepared for high demand tutoring classes by attending continuous education sessions.- College Reading & Learning Association Certified Tutor | September 2021 - Current

Problem: Students, especially new ones, require additional support to find success in their classes for a variety of reasons. MSOE does provide tutoring services, but for struggling students, that may not be enough. My role as a residential tutor is to provide a more convenient and more frequent tutoring offering for students in the residence halls. New students who need the most support also live in the residence halls, and I work with Residence Life concurrently as a Resident Assistant to ensure that residents are obtaining the support they need.- Worked with Residence Life to leverage my tutoring skills and provide a convenient tutoring option in the residence halls.- Individually tutored ~10 residents a week over two two-hour-long sessions in mathematics, computer science, and technical communication classes.- Acted as a resource for Residence Life concurrently with my Resident Assistant position.

Problem: Students living in the residence halls on campus, especally new students, need academic, social, and professional support as they transition into their adult lives. Resident assistants fill the gap by living in the halls with residents and acting as a resource for their needs.- Developed a sense of community in the residence halls where residents could grow academically, socially, and professionally.- Hosted 14 events per year that residents enjoyed with each other and found helpful for their wellbeing.- Resolved conflicts with residents in order to create a safe environment for them and others in the residence halls.

Problem: DigiKey is a five billion dollar company that currently offers over 4,000,000 products and their corresponding value-add services to its customers. With such a large catalog, it can be difficult for customers, large and small, to find the parts they need, especially if they are not completely confident in their bill of materials. To fill that gap, I individually ported a legacy recommendation engine to make recommendations for customers’ bill of materials that can be used in a variety of situations (e.g. checkout, cart abandonment, and order confirmation recommendations).- Individually migrated legacy bill of materials real-time recommendation API from Redis to Dataiku and Snowflake for use in online catalog.- Implemented backend pairwise Pearson correlation recommendation algorithm in Snowflake SQL to improve runtime performance.- Developed frontend Python REST API to serve batched recommendations.- Benchmarked Snowflake’s performance for viability in real-time recommendation systems.- Recommendation model led to ~10-20x increase in order processing efficiencyProblem: The loading of tape and reel units for processing in DigiKey distribution centers is currently a manual process prone to human error. The process itself is not currently being targeted for automation, but aspects of it could be automated to reduce the loading error rate. In particular, if a reel is loaded incorrectly (e.g. backwards or unlocked), processing speed could be greatly affected. To help alleviate the problem, I worked with a team of three to develop a machine learning algorithm to determine whether a reel was loaded properly given an image of it.- Developed Dataiku classification model and API using Postman with team of three to automate tape and reel loading in DigiKey distribution centers.- Developed novel algorithm to segment and straighten barcodes in images using contour analysis for previous system.

Problem: Circular saw blades (CSBs) undergo a large variety of quality assurance testing to ensure they can beat our competitors. Analysis of the test results for a given blade is a painstaking task, as an engineer in the lab must manually inspect every tooth on it, which can range from 24 to 100. To fill this gap, I worked with a team of four to develop a fixture to automate the process.- Developed fixture with team to automate CSB wear analysis and accelerate departmental CSB iteration times.Created GUI for the fixture that integrated with image capturing hardware and the Microsoft Azure (Azure) platform.- Created image processing and Azure CustomVision machine learning models to clean and classify the fixture’s CSB images.- Refined fixture requirements through cross-functional communication regarding blade design and the Azure platform.- Eliminated tedious manual labor for engineers, leading to ~24x speedup in CSB analysis.

Problem: The reporting services we provided to our end users was a very manual process prone to errors and misunderstandings, leading to long turnaround times and unhappy customers. To fill that gap, I worked with a team of four to improve and automate the reporting and report development processes.- Developed IBM Cognos Analytics 11 (Cognos) reports to advance self-service for DSSI ecommerce platform customers.- Created the first GitLab Java CI/CD pipeline ever at DS, automating development with Amazon Web Services (AWS) through Terraform.- Eliminated tedious manual labor for four teammates, leading to a 10% increase in ticket response times to meet KPIs.- Participated in internal hackathons, including 1st place in AWS DeepRacer competitions and panda launcher assembly.

(Also at St. Francis, WI)Led development, implementation, and testing of a RoboRIO and Raspberry Pi embedded system in Java and vision tracking (cut short due to COVID-19)Helped develop, implement, and test a RoboRIO and Raspberry Pi embedded system in C++ and Pixy and Chameleon vision trackingTaught interested team members’ programming skills through in-person teaching, training videos, and hands-on exercisesCreated methods to ensure the team’s robot fit within competition constraints

(Also at St. Francis, WI)Chairman’s essay and presentation leaderPursued other awards teams could win

- Flagged motorcycles for parking during Summerfest’s operation.- Provided customer service to those around the venue.