From a young age, liberal arts and engineering graduate Avi Peltz has been intrigued by artificial intelligence (AI) and machine learning (ML). AI involves computer systems attempting to model and apply human-like intelligence, while ML, a branch of AI, focuses on using data and algorithms to replicate human learning.
“Machine learning was this shiny, cool new thing that seemed really powerful,” Peltz said. “I have always been curious about it.”
Growing up in tech-savvy Berkeley, Peltz watched many demonstrations of computer vision models that could identify objects and images. He described it as “astounding.” He saw developers bringing “incredible” creations to life and aspired to be a part of that.
Peltz has always been technically minded, he said. Throughout high school, he built different websites and computer games. He also worked on various robotic and ML projects. This quickly taught him how ineffective pre-existing tools were that are intended to help perform basic tasks.
After growing irritated with the complexity of ML tools, he sought to create a solution. That is when he began building his startup, TensorMaker.
TensorMaker is a platform that makes building ML applications fast, easy and accessible to everyday developers.
“A lot of the impetus of wanting to build TensorMaker was my own frustrating process during ML projects and not having very helpful tools,” Peltz said. “I wanted to create a better user experience.”
TensorMaker aims to streamline the ML pipeline — a series of steps including the development, deployment and monitoring of an ML model. TensorMaker guides users through this pipeline without needing any prior ML experience.
“Most people trying to build these types of tools are focusing on enterprise ML teams and making them more productive,” Peltz explained. “We’re focusing on everyday developers.”
By making his technology accessible, Peltz is enabling technology companies in domains ranging from agriculture, manufacturing and life sciences to take advantage of the power of ML while still maintaining their focus on the core competency of their business.
After working on TensorMaker throughout college, he decided to compete in the Cal Poly Center for Innovation and Entrepreneurship (CIE) Innovation Quest (iQ).
Innovation Quest is a competition that encourages student entrepreneurs to pursue their innovative ideas and help with the funding and resources needed to launch their ventures.
TensorMaker was one of 14 finalists to pitch their startup at iQ 2023. Although he did not win the competition, iQ allowed him to develop more answers about his business, talk to potential users and do customer research, Peltz said.
Following Innovation Quest, Peltz applied for the CIE Summer Accelerator. The Summer Accelerator is a program for Cal Poly students and recent alumni that provides them with mentorship, weekly workshops and $10,000 in seed funding.
TensorMaker was one of the eight teams accepted to the program.
“We received our first customer through the CIE community and without being here, we probably would not have had that relationship,” Peltz explained.
The CIE provided Peltz with not only the business language that he wasn’t initially confident in but also mentors who continue to lead him in the right direction, he said.
“Being in a space with people who are motivated to work on exciting projects motivates you to work even more,” Peltz said.
Over the course of the Summer Accelerator, TensorMaker worked on releasing its initial prototype. Peltz is eager to make his technology accessible to everyday developers and to explore the potential impacts of TensorMaker on a broader scale.
“Whether it’s making agriculture more efficient and measured or being able to identify tumors in radiology scans, there are so many applications that can use these ML techniques and improve human well-being on the planet,” Peltz said.
TensorMaker, along with the rest of the 2023 Summer Accelerator cohort, will pitch their startup and showcase their progress at Demo Day on Sept. 8 at 4 p.m. at SLO Brew Rock. Tickets are available here.