Brains: The Sweet Spot for Blood Sugar Tracking with January AI CEO Noosheen Hashemi

I remember the first time I heard about diabetes. I was in elementary school, frantically searching for causes and symptoms, desperately trying to figure out if it was a death sentence. My heart was racing. I didn’t have a lot of time alone on the computer, and I knew that Googling phrases like “life expectancy with diabetes” was something my parents would have questions about if they caught me. But I couldn’t help it; Nick Jonas had just been diagnosed with type 1 diabetes, and I was supposed to go to his concert soon. This wasn’t just any concert. It was my first concert ever, and Nick was my Jonas Brother, the one whose teen magazine spreads plastered my walls and whose face was on my second-favorite T-shirt (only topped by the one my mom let me get from Hot Topic that said, “I was a vegetarian before it was cool”). My research left me more confused, but Nick ended up performing, and my shirt looked great at the concert.

He even capitalized on the diagnosis and did some influencer marketing around diabetes. While Nick was able to keep being a boy band heartthrob, his diabetes didn’t disappear, and the disease, along with those impacted by it, has stayed increasingly on my radar.

@nickjonas

WOW… 😎 the all-new @dexcom G7 is available tomorrow in the US for people with all types of diabetes. #Sponsored #DexcomMagicMoment #Diabet... See more

Flash forward to a year ago, and another man I see on the internet far too much was talking about his blood sugar: Bryan Johnson. In a clip, he explained how his blood glucose levels were ideal, which he figured out after wearing an uncomfortable blood glucose monitor for ten days. As someone who has avoided going to the ER purely because there’s a chance they might hook me up to an IV, a CGM sounded like my worst nightmare.

It’s not the needle that’s the problem. At one point in college, I had eight ear piercings, plus my nose and belly button pierced. I looked like I was trying to attract a murder of crows, and I wasn’t afraid of needles. It’s having the needle stay in my body that I don’t like. As the year went on, I saw more and more people from my side of the tech bro podcast world talk about blood sugar and managing it; from Diary of a CEO to Tim Ferriss, and even some X discourse.

Even beyond the corners of my chronically online side of the internet, consumers are seeing brands like Oura introduce a feature that tracks blood sugar levels through a partnership with Dexcom, alongside whispers of Apple developing a non-invasive blood glucose monitoring system for the Apple Watch. In a few years' time, I expect this trend of tracking your blood sugar to snowball into something as common as monitoring the hours you sleep.

This brings me to my conversation with Noosheen Hashemi, founder and CEO of January AI, a health tech company focused on managing chronic diseases. In 2024, January launched an app that uses AI to predict how food will affect your blood sugar, with no sensors, needles, or even eating required. It’s a huge step in tackling issues like diabetes, prediabetes, and obesity. Before founding January, Noosheen was a VP at Oracle, where she helped grow the company’s revenue from $26 million to $3 billion a year.

I had dinner with Noosheen and used the app with her; her enthusiasm for the product is unmatched. I had the pleasure of speaking with her again for Braun & Brains, diving deeper into the company’s mission and why she thinks this is a problem worth solving.

Noosheen Hashemi is the founder and CEO of January AI, a company that leverages artificial intelligence to advance precision health and preventative care, with a focus on metabolic health.

Rachel: Some people think metabolic health is just a passing trend. Do you believe this focus is here to stay?

Noosheen: No, I think this is definitely going to stick around. We began putting CGMs on healthy people in 2017, and a lot of companies followed. Of course, we are the only AI company of the bunch. We always intended to leverage the power of hardware to provide AI predictions so people don’t have to wear hardware all the time.

Our vision was to help people use hardware to see where they are and get educated, but also to have tools for those who can’t afford hardware or find it too invasive. In our first and second product versions, we offered a CGM. In our third version, available in the app store now, it works without sensors, expanding access so more people can learn about metabolic health. If they want to use a sensor, they can, but they don’t have to. I actually think the moment for blood sugar awareness in America is still a few years away, probably a 10-year cycle from when we started putting CGMs on healthy people.

Rachel: Many online discussions suggest that non-diabetics don’t need continuous glucose monitors. Why should they still consider tracking their blood sugar?

Noosheen: I agree that if they are not insulin-intense, they don’t need to wear a sensor. But a lot of people don’t know they have prediabetes or diabetes. Twenty-two percent of people believed to have diabetes in this country are undiagnosed, and 90 percent of people with prediabetes are undiagnosed. People don’t really know, so it’s kind of like Russian roulette.

In the U.S., we don’t screen for metabolic disease unless someone is visibly obese and over 40. If people want to know and can afford it, they can wear a sensor for 10-14 days and see how they’re really doing. If they don’t want something invasive or expensive, they can download January AI and start learning by just taking pictures of their food. The AI does the rest.

Rachel: When you showed me your technology over dinner, I was amazed. How did you develop this system without requiring physical glucose monitors?

Noosheen: It seems magical, but it’s seven and a half years of research. It’s millions of data points from thousands of people who wore sensors. We’ve been published twice, most recently in November 2023 and January 2025 in NPJ Digital Medicine, for our AI’s ability to scale weight loss and behavior change.

We developed our own generative AI that predicts glucose levels. It has taken tens of millions of dollars and a lot of AI talent to break down food into a measurable assay. We built a nutritional inference model, created glycemic index and glycemic load calculations, and developed a food database with 54 million foods. The app allows users to take pictures of meals, and the AI identifies the food, breaks it down into ingredients, and predicts how it will affect blood sugar. Users can then modify portions and see how changes might improve their response.

Rachel: What has been the most difficult part of building January AI?

Noosheen: Early on, the challenge was hiring AI talent. In 2017, people wanted to work at OpenAI or on tools for everyone rather than health applications. Another challenge was having no experience building a tool like this. We had to run clinical trials just to observe people and see what would happen, which was expensive and time-consuming.

Our early products were clunky because AI engineers develop models in Jupyter notebooks, which can’t just be dropped into an app. We had to build all the data engineering infrastructure to make it work smoothly. Over time, we got our infrastructure right so R&D could seamlessly transition into an app experience.

Now, a challenge is that while our predictions could work for Type 1 diabetics, we don’t want to go through the regulatory process. It would require an expensive, large trial. Also, many people still don’t fully trust AI or expect to pay for it. And blood sugar awareness in the U.S. is still developing. People wake up wanting to lose weight, but they don’t yet think about monitoring their blood sugar as a daily priority.

Rachel: What major health trends do you anticipate in the next five years, and how will January AI fit into them?

Noosheen: Metabolic health is an unstoppable trend. You saw Oura buy Vari, and Oura will have metabolic health features soon. Apple will follow. Others will too. Metabolic health will become consumerized, and I think everyone will be tracking their blood sugar in the future.

The omics revolution is another big trend. Managing health requires more than just knowing cholesterol levels. People will care about body composition, get DEXA scans, and wear CGM sensors from companies like Lingo or Stelo. There will be advancements in proteomics too. Wearables, omics, and metabolic health will dominate.

I also think we will see major changes in healthcare under this administration. I expect serious efforts to make electronic health records interoperable and to improve data accessibility for patients and doctors. AI will become central to integrating all this data, leading to better health insights.

Rachel: You believe the current administration will improve healthcare, but my brother, who is applying to med schools right now, is concerned the opposite will happen. What changes do you expect?

Noosheen: Healthcare has a lot of challenges. Regulations meant to protect privacy sometimes maintain monopolies and keep pricing opaque. Insurance companies and healthcare providers have secret deals that make costs unpredictable for patients. People already pay huge premiums, then get surprise bills.

I hope we will see more transparency, removal of unnecessary regulations, and better data integration so AI can help people understand their health. In finance, insider trading and self-dealing are tightly controlled. Healthcare should have the same level of transparency so consumers can make informed choices.

Discussion about this podcast