Birago Jones is the CEO and Co-Founding father of Pienso, a no-code/low-code platform for enterprises to coach and deploy AI fashions with out the necessity for superior information science or programming expertise. At the moment, Birago’s clients embody the US authorities and Sky, the most important broadcaster within the UK. Pienso relies on Birago’s analysis from the Massachusetts Institute of Expertise (MIT), the place he and his co-founder Karthik Dinakar served as analysis assistants within the MIT Media Lab. He’s a distinguished authority within the intersection of synthetic intelligence (AI) and human-computer interplay (HCI), and an advocate for accountable AI.
Pienso‘s interactive learning interface is designed to enable users to harness AI to its fullest potential without any coding. The platform guides users through the process of training and deploying large language models (LLMs) that are imprinted with their expertise and fine-tuned to answer their specific questions.
What initially attracted you to pursue your studies in AI, HCI (Human Computer Interaction) and user experience?
I had already been developing personal projects focused on creating accessibility tools and applications for the blind, such as a haptic digital braille reader using a smartphone and an indoor wayfinding system (digital cane). I believed AI could enhance and support these efforts.
Pienso was initially conceived during your time at MIT, how did the concept of training machine learning models to be accessible to non-technical users originate?
My co-founder Karthik and I met in grad school while we were both conducting research in the MIT Media Lab. We had teamed up for a class project to build a tool that would help social media platforms moderate and flag bullying content. The tool was gaining lots of traction, and we were even invited to the White House to give a demonstration of the technology during a cyberbullying summit.
There was just one problem: while the model itself worked the way it was supposed to, it wasn’t skilled on the fitting information, so it wasn’t in a position to determine dangerous content material that used teenage slang. Karthik and I had been working collectively to determine an answer, and we later realized that we may repair this concern if we discovered a method for youngsters to instantly practice the mannequin information.
This was the “Aha” second that might later encourage Pienso: subject-matter consultants, not AI engineers like us, ought to have the ability to extra simply present enter on mannequin coaching information. We ended up growing point-and-click instruments that enable non-experts to coach giant quantities of knowledge at scale. We then took this expertise to native Cambridge, Massachusetts faculties and elicited the assistance of native youngsters to coach their algorithms, which allowed us to seize extra nuance within the algorithms than beforehand doable. With this expertise, we went to work with organizations like MTV and Brigham and Ladies’s Hospital.
Might you share the genesis story of how Pienso was then spun out of MIT into its personal firm?
We at all times knew that this expertise may present worth past the use case we constructed, however it wasn’t till 2016 that we lastly made the leap to commercialize it, when Karthik accomplished his PhD. By that point, deep studying was exploding in recognition, however it was primarily AI engineers who had been placing it to make use of as a result of no one else had the experience to coach and serve these fashions.
What are the important thing improvements and algorithms that allow Pienso’s no-code interface for constructing AI fashions? How does Pienso be certain that area consultants, with out technical background, can successfully practice AI fashions?
Pienso eliminates the boundaries of “MLOps” — information cleansing, information labeling, mannequin coaching and deployment. Our platform makes use of a semi-supervised machine studying method, which permits customers to begin with unlabeled coaching information after which use human experience to annotate giant volumes of textual content information quickly and precisely with out having to jot down any code. This course of trains deep studying fashions that are able to precisely classifying and producing new textual content.
How does Pienso supply customization in AI mannequin improvement to cater to the particular wants of various organizations?
We’re sturdy believers that nobody mannequin can clear up each downside for each firm. We want to have the ability to construct and practice customized fashions if we wish AI to know the nuances of every particular firm and use case. That’s why Pienso makes it doable to coach fashions instantly on a corporation’s personal information. This alleviates the privateness issues of utilizing foundational fashions, and may also ship extra correct insights.
Pienso additionally integrates with current enterprise programs via APIs, permitting inference outcomes to be delivered in several codecs. Pienso may also function with out counting on third-party companies or APIs, that means that information by no means must be transmitted exterior of a safe atmosphere. It may be deployed on main cloud suppliers in addition to on-premise, making it an excellent match for industries that require sturdy safety and compliance practices, resembling authorities businesses or finance.
How do you see the platform evolving within the subsequent few years?
Within the subsequent few years, Pienso will proceed to evolve by specializing in even better scalability and effectivity. Because the demand for high-volume textual content analytics grows, we’ll improve our potential to deal with bigger datasets with sooner inference occasions and extra advanced evaluation. We’re additionally dedicated to decreasing the prices related to scaling giant language fashions to make sure enterprises get worth with out compromising on pace or accuracy.
We’ll additionally push additional into democratizing AI. Pienso is already a no-code/low-code platform, however we envision increasing the accessibility of our instruments much more. We’ll repeatedly refine our interface so {that a} broader vary of customers, from enterprise analysts to technical groups, can proceed to coach, tune, and deploy fashions with no need deep technical experience.
As we work with extra clients throughout various industries, Pienso will adapt to supply extra tailor-made options. Whether or not it’s finance, healthcare, or authorities, our platform will evolve to include industry-specific templates and modules to assist customers fine-tune their fashions extra successfully for his or her particular use instances.
Pienso will change into much more built-in inside the broader AI ecosystem, seamlessly working alongside the options / instruments from the main cloud suppliers and on-premise options. We’ll deal with constructing stronger integrations with different information platforms and instruments, enabling a extra cohesive AI workflow that matches into current enterprise tech stacks.
Thanks for the nice interview, readers who want to study extra ought to go to Pienso.
