Humane Science in Action: How a Canadian Biotech Company Is Creating Non-Invasive Tools for Early Disease Detection

A conversation with Eric Ocelewski, Co-Founder and CEO of Biotech Company Pi and Power, and member of the Animal Alliance Advisory Group on Humane Science. 

In this interview, we explore Eric’s personal “why,” the science behind Pi and Power’s technology, and how this model of real-world, non-animal data generation can help shift the biomedical field toward more ethical innovation.  

Read on to learn more. 

Can you start by introducing Pi and Power?  

Pi and Power is building humane, non-invasive biosensing tools that treat “early” as a design spec, not an afterthought. We use radiofrequency and metamaterial sensors, coupled to AI, to read subtle changes in tissue – things like infection, stress, and metabolic state – without needles, reagents, or sending samples to a lab. 

We are starting in veterinary medicine, because companion animals are severely underserved on both the diagnostics and ethics fronts, and because the physics is species-agnostic. The same platform that helps a dog with cancer or diabetes can later be translated into human applications with proper calibration, data, and regulation. 

What inspired the creation of the company, and what core problem are you trying to solve? 

Pi and Power came out of frustration with how late and how invasive our current system is. As a patient, and as someone who has watched a beloved dog get a terminal diagnosis, you see the same pattern: we wait until there is obvious structural damage before we act. By then, the options are brutal, expensive, and often palliative. 

The core problem we are attacking is the “temporal gap” between when a disease becomes physically detectable and when it becomes clinically obvious. Biology leaves signatures in fields and waves long before visible damage appears. If we can reliably read those signatures non-invasively, then “early” becomes normal, not exceptional. 

The second problem is ethical: too much of our innovation pipeline is built on animals that will never benefit from the treatments they help create. I do not accept that as the only way to do serious science.

Tell us about your background. What academic or research experiences brought you into this space, and how have they influenced your approach in science, particularly humane science? 

My background is in biophysics and advanced mathematics, with a lot of time spent thinking about how fields, waves, and information flow through living systems. I have modeled everything from antimicrobial peptides using probabilistic methods to differentiating viruses and oligonucleotides using microwave sensors and complementary split-ring resonators. 

That training did two things. First, it convinced me that biology is not just chemistry in a bag – it is also an electromagnetic and information system. Second, it gave me a respect for rigor that co-exists with imagination. Humane science, in my view, is not “softer” science. It simply holds itself to two constraints at once: technical validity and moral coherence. 

Being a patient and a caregiver layered something else on top: a very low tolerance for beautiful theories that never arrive in the clinic, and for protocols that are convenient for institutions but brutal for animals and families. 

What is your personal “why”? How did your professional or personal experience shape your commitment to creating technologies that benefit both animal and human health? 

My “why” is very simple: I do not want the next generation to go through what mine has gone through with the same crude tools and the same late timing. 

I have had cancer and major surgery. I am immunocompromised. I have watched my dog – a family member, not a “pet” in the disposable sense – receive a two-month prognosis for osteosarcoma. You start asking very direct questions: 

• Why did we only see this now? 
  
• Why is “early” always a lucky accident rather than the default? 
  
• Why does the path to progress still run through animals in cages that will never see a home, while my own dog can’t access cutting edge diagnostics? 

Professionally, I have seen how much signal we throw away in routine care. Every visit is an opportunity to learn from real animals and real people in real environments, if you build the tools and the data infrastructure correctly. That is why Pi and Power is designed around “companion animals first” and “shared benefit” if a study does not help the actual patient in front of us, I am not interested. 

How does your technology work? For people unfamiliar with biomedical devices, how would you explain what your device does and the potential impact it can have? Feel free to explain at a higher scientific level as well. 

For a non-technical explanation: 
 
You can think of our device as a very sensitive “radio mirror” for the body. We send tiny, safe radiofrequency signals into or near tissue and measure how they bounce, bend, and slow down. Healthy tissue, inflamed tissue, cancerous tissue, high blood sugar, stress – they all change the way those waves behave in subtle but detectable ways. 

Our system captures those patterns and feeds them into AI models that have been trained on paired data – the wave signatures plus conventional lab values and clinical outcomes. Over time, the device learns to say things like: “this pattern looks like rising cortisol,” or “this signature is consistent with early infection,” without drawing blood every time. 

At a higher scientific level: 

• We use metamaterial-based sensors, including complementary split-ring resonators (CSRRs), coupled to a vector network analyzer operating in the 10 MHz to several GHz range. 

• We measure complex S-parameters (S11, S21, etc.) across thousands of frequency points, effectively sampling the dielectric and dispersive properties of tissue. 

• These high-dimensional signatures are modeled using modern machine learning and uncertainty-aware methods to map EM response patterns to biochemical and physiological states such as glucose, inflammatory markers, stress hormones, and potentially even viral or bacterial loads. 

The impact, if we get this right, is a shift from episodic, invasive, late diagnostics to continuous, non-invasive, early-state monitoring – first for animals, then for humans. 

Your testing model is unique. You work with consenting pet owners through veterinary partnership(s) rather than using animals in invasive research. Was this always the plan for the testing phase? What ethical considerations guided this approach? 

Yes, this was the plan from the outset. I was never interested in building a company whose progress depends on creating suffering upstream so we can claim to reduce suffering downstream. 

Our model is: 

• Work with veterinary hospitals where animals are already receiving care for conditions they genuinely have. 

• Offer non-invasive scans that add information but do not add harm. 

• Give something back to the animal and the owner in the form of better monitoring, earlier warnings, or finer-grained tracking of treatment response. 

The key ethical considerations are: 

• No induced disease. We do not “create” illness to test our device. 

• No terminal-use colonies. The animals we study go home with their families. 

• Informed consent from owners, with clear communication about what we know and what we are still learning. 

• Data as a shared asset – the goal is to build models that benefit the wider animal and human population, not just our balance sheet. 

If a study design requires an animal to suffer purely for data, we do not do that study. Period. 

From your perspective, what are the limitations of traditional animal research methods? And how do you see the scientific landscape evolving toward more human-relevant, non-animal methods? 

Traditional animal research has given us important insights, but it has hard limits: 

• Translation gap: Rodent models in artificial environments often fail to predict human responses, especially for complex, chronic, multi-factorial diseases. 

• Context loss: Real patients – human or animal – live in rich environments with stress, diet, microbiome, and co-morbidities. You cannot reproduce that in a cage with a standardized diet and artificial light cycles. 

• Ethical cost: The moral calculus assumes a high probability of benefit to others, but the actual translation rate from many animal models to approved therapies is low. That is a problem. 

• Inertia: Many protocols persist because “this is how we have always done it,” not because they are still the best way to answer the scientific question. 

The landscape is already shifting. We see growth in: 

• Human organoids and organ-on-chip systems. 

• In silico models and AI-driven drug and toxicity prediction. 

• Advanced imaging and biosensing that can read biology in vivo without destruction. 

• Real-world evidence from clinics, wearables, and at-home devices. 

Our work sits in that last category: non-invasive, in-clinic biosensing in real animals with real diseases, under veterinary care. It is messy, but it is honest and human-relevant. 

What role do you think companies like yours can play in shifting the biomedical field toward more compassionate, human-relevant science? 

We have to prove that humane science is not just morally preferable, but technically superior and economically viable. If all we do is make people feel bad about the status quo, nothing changes. 

Companies like Pi and Power can: 

• Build diagnostic platforms that never required animal colonies in the first place. 

• Publish and share methods that show regulators and clinicians that non-animal, real-world data can meet or exceed traditional benchmarks. 

• Collaborate with veterinary partners so that animals are early beneficiaries, not late afterthoughts. 

• Make open or low-friction toolchains so others can replicate and extend what we do without reinventing the wheel – particularly academic and non-profit groups. 

If we can demonstrate that you can de-risk a device, run meaningful studies, and get to market while aligning with animal welfare and human relevance, then “humane-first” becomes a credible default, not a niche. 

What’s next for Pi and Power? What future applications, studies, or collaborations are you most excited about? 

Near-term, we are focused on three things: 

1. Veterinary pilots 

• Rolling out structured pilot studies with our veterinary partners in Toronto and beyond, starting with use cases like non-invasive glucose monitoring, infection detection, and cancer monitoring in dogs. 

• Building longitudinal datasets that track animals over weeks and months, not just single snapshots. 

2. Multi-modal sensing and AI 

• Extending beyond microwaves to integrate bioelectric, optical, and possibly acoustic signals, so we can see a richer picture of tissue state. 

• Refining our AI pipeline to be uncertainty-aware, so the device can say “I don’t know” when it should, and to make the models more interpretable to clinicians. 

3. Translational bridge to human health 

• Designing early human feasibility work where appropriate – for example, non-invasive stress and metabolic monitoring – while staying inside regulatory and ethical guardrails. 

• Exploring collaborations with hospitals, aging research groups, and bioelectric labs to investigate how these field-based signatures might be used for earlier intervention, not only in disease, but potentially in aging trajectories. 

Longer-term, I am excited about a world where: 

• A veterinarian or physician can get a rich, non-invasive “field profile” of a patient in minutes. 

• Companion animals help us validate humane, early diagnostics that then translate into better human care. 

• The default question becomes “how early can we see this?” rather than “how aggressive can we be now that it is late?” 

That is what Pi and Power exists to build. 

Learn more about Pi and Power on their website and/or connect with Eric on LinkedIn here.