OxIgnite Ep7: Sahil Agarwal, Enkrypt AI
Learn how Enkrypt AI enables secure on-premises deployment of advanced AI models, safeguarding data privacy and intellectual property for enterprises.
Welcome to the OxIgnite Podcast hosted by Yusuf Raza and Michael Hutson - two entrepreneurial MBA students at the University of Oxford. Together, we navigate the dynamic journeys of Oxford-linked founders, venture capitalists, and trailblazing innovators. Our mission? To fuel the fire in Oxford's next generation of boundary-pushing change-makers.
Our seventh episode features Sahil Agarwal, the founder of Enkrypt AI, which provides a robust and secure platform for on-premises deployment of advanced AI models, ensuring data privacy and intellectual property protection for enterprises.
For investors: Enkrypt AI is currently a part of the UC Berkeley Skydeck accelerator program and will be looking to raise funds after the Skydeck Demo day in September
Discover how Enkrypt AI's cutting-edge solutions enable AI model providers to securely deploy their advanced models in complex environments, offering newfound revenue streams and protecting precious intellectual property.
Data security is emerging as the cornerstone for commercializing machine learning technologies, protecting proprietary algorithms, and fostering enterprise interest in adopting AI.
With the AI revolution sweeping the world, Enkrypt AI provides a state-of-the-art solution that protects proprietary algorithms and models.
As AI becomes an integral part of every human and business workflow, Enkrypt AI's dedication to securing AI model deployment paves the way for widespread adoption.
Sahil also turns the tables on the host Yusuf Raza, and asks him about his entrepreneurship journey with his startup Panalyt.
Join us for an engaging episode filled with valuable insights about the challenges that AI model providers and enterprises face with deploying AI models.
Get ready, get set, get OxIgnited!
List of Topics
Introduction to Sahil Agarwal, founder of Enkrypt AI, and an overview of the company's core mission in the AI industry.
Enkrypt AI's focus on providing a secure platform for AI model providers, enabling deployment in complex customer environments.
Sahil Agarwal's impressive background in applied mathematics, passion for solving real-world problems, and expertise in applied machine learning.
Enkrypt AI's mission to address the challenge of monetizing AI models while ensuring data privacy and security in on-premises deployments.
Identifying the potential clientele for Enkrypt AI, including major AI model providers and various companies in different industries.
Sahil Agarwal's journey from applied mathematician to AI entrepreneur, fuelled by a passion for AI and real-world impact.
Challenges faced during Enkrypt AI's startup inception, including funding scarcity and market sentiment during an economic downturn.
The importance of customer discovery and understanding needs for building a successful and sustainable startup.
Success factors, including participation in the Berkeley Skydeck accelerator program and partnerships with consulting companies.
The profound impact of the AI hype on Enkrypt AI and the future of AI, emphasizing data security and privacy in commercializing machine learning technologies.
We also cover OxIgnite host Yusuf Raza’s entrepreneurship journey with his Startup Panalyt.
Summary
Our seventh episode features Sahil Agarwal, the founder of Enkrypt AI, which provides a robust and secure platform for on-premises deployment of advanced AI models, ensuring data privacy and intellectual property protection for enterprises.
Enkrypt AI's primary focus is to offer a robust and secure platform for AI model providers, allowing them to deploy their advanced AI models in complex customer environments with utmost confidence.
This unique capability empowers Enkrypt AI's clients to not only boost their revenue streams but also safeguard their valuable intellectual property while deploying AI models precisely where their customers need them.
Enkrypt AI Founder Sahil Agarwal has a fascinating background.
Armed with a profound understanding of applied mathematics, Sahil is driven by a passion for solving intricate real-world problems that can make a tangible difference in people's lives. His expertise lies in applied machine learning, optimizing human technology, and advocating for the adoption of clean technologies.
Sahil has contributed to over ten publications in prestigious international journals and secured three valuable patents. His groundbreaking research work has garnered recognition in renowned publications like the American Physical Society, Yale Scientific Magazine, Yale Daily Newspaper, and Yale News.
Sahil shares more about Enkrypt AI's mission, seeking to understand the core problem the company addresses in the AI industry.
He enthusiastically explains that while AI models, particularly large language models, demand significant investments in terms of time and millions of dollars during their development and training phases, monetizing these models can be a daunting challenge.
The primary reason behind this is that enterprises place paramount importance on data privacy and security, and they are unwilling to take any risks that might compromise their data integrity and customer trust.
To meet the stringent requirements of enterprises, on-premises deployment of AI models has become the preferred choice over using APIs, which may pose data privacy risks.
This crucial challenge is where Enkrypt AI comes to the rescue.
The company offers state-of-the-art solutions that cater to the rising demand from enterprises for on-premises deployment while ensuring that AI models remain secure, protected, and tamper-proof.
By doing so, Enkrypt AI opens up new and lucrative revenue streams for AI model providers, enabling them to effectively sell their advanced models to enterprises with confidence and peace of mind.
Sahil further elaborates that while major players like OpenAI, Anthropic, Stability, and Cohere are well-known AI model providers, there are numerous other companies working in diverse industries such as medical devices, gaming, and more that also offer proprietary AI models. These models, whether built from scratch or fine-tuned from open-source models, represent the potential clientele for Enkrypt AI.
Speaking about his transition from working on Applied Mathematics in academia to working on AI in the industry, Sahil talks about his time at Accrete AI, a fintech company where he led the AI team.
During this time, Sahil's mathematical background proved immensely valuable in developing cutting-edge data algorithms, enabling the understanding of noise characteristics and other key attributes of the data.
Additionally, he effectively applied machine learning algorithms to detect exoplanets in vast data sets collected from satellites and telescopes. This successful application of machine learning in astronomy ignited Sahil's passion for exploring AI further.
Upon realizing his fervor for entrepreneurship and the potential to make a substantial impact, Sahil embarked on his journey to found Enkrypt AI.
Recognizing that the best time to start a venture is as soon as possible, Sahil embraced the challenge of building a startup that addresses critical industry needs with innovative solutions.
Enkrypt AI's commitment to providing secure AI model deployment for enterprises is driven by a potent combination of mathematical insights and cutting-edge AI solutions, fostering innovation, and addressing real-world challenges in the ever-evolving AI industry.
Sahil shares valuable insights and experiences from his entrepreneurship journey.
He emphasizes the importance of customer discovery and understanding their pain points before developing a solution.
Coming from an academic background, Sahil acknowledges that one common mistake in the industry is being fixated on the technology or solution and trying to force it onto a problem.
Instead, he advises entrepreneurs to prioritize understanding the customer's needs and the urgency of the problem they want to solve.
Conviction in these three aspects—problem, solution, and customer need—is crucial for building a successful and sustainable startup.
Sahil also discussed the major challenges faced during their Enkypt AI's inception.
The timing of their venture coincided with recession fears, making funding scarce. To overcome this, Sahil's team decided to bootstrap the startup.
However, the economic downturn also affected their potential customers, who were reluctant to engage with startups and allocate budgets for new projects.
Sahil explains how they dealt with these challenges by finding the right customers through networking.
They secured two pilot projects, but progress was slow due to the cautious market sentiment. To adapt to the situation, they realized that bringing data to models and vice versa was a challenging process. As a result, they shifted their approach to model security, enabling model providers to ship their models to the data's location.
Moving forward, Sahil anticipates operational challenges regarding team size and experience. However, they have already made significant strides in this area, hiring four excellent individuals from renowned institutions and engaging three advisors with impressive industry backgrounds.
One of the significant factors contributing to their success has been their participation in the Berkeley Skydeck accelerator program. Sahil praises the program's value, particularly its access to a vast network of experts, advisors, and prospective customers. This network has helped them secure pilots, proofs of concept, and paying customers, which are crucial for any startup's growth.
Additionally, Sahil shares how partnerships with consulting companies have provided their startup with credibility. Seeing these partners involved with their company makes prospective clients trust them more, leading to an increase in interest.
Sahil also delves into the profound impact of the AI hype, fueled by models like GPT-3, on their startup, Enkrypt AI.
He compares the current AI revolution to the explosion of the internet in 1993, predicting that AI will become an integral part of every human and business workflow in the next five to seven years. This transformation creates a significant opportunity for businesses to leverage AI, but it also brings forth critical challenges related to data security and privacy.
Sahil emphasizes that to make AI a reality for businesses and ensure the commercialization of machine learning technologies, data security is the key pillar. As AI becomes more pervasive, the need to protect proprietary algorithms and models becomes paramount.
Discussing the aftermath of the leak of Samsung's proprietary code base to OpenAI, Sahil points out the repercussions for various industries that handle sensitive and confidential data, such as financial institutions, healthcare providers, and semiconductor companies. The incident prompted organizations to exercise caution and restrict the usage of chat GPT and similar models to safeguard their data and intellectual property.
Regarding future growth plans, Sahil reveals that the Berkeley Skydeck accelerator program is ongoing, with a demo day scheduled for September.
At the moment, their primary focus remains on serving their customers, but they are preparing to enter investor fundraising mode in mid-August.
The company aims to raise a significant funding round by the end of September to support its expansion and innovation.
Shifting the conversation to a more personal note, Yusuf inquires about Sahil's approach to managing stress as a startup founder and his perspective on wellness.
Sahil attributes his ability to handle stress to the strong support he receives from his family. Additionally, he finds solace and relaxation in engaging in sports, particularly racket sports like tennis, squash, and table tennis. These activities allow him to disconnect from the demanding startup environment and recharge his energy and focus.
Sahil can be contacted by listeners via his email or LinkedIn for further discussions or inquiries. *Links shared below*
Bonus Content: In this episode, our guest Sahil Agarwal also turns the tables on OxIgnite Host Yusuf Raza and asks him to share more about his entrepreneurial journey with his startup Panalyt.
Yusuf narrates a hackathon experience during university where he and a friend discovered discrepancies in flight prices on Skyscanner based on different countries' versions of the website. This led to him reaching out to Skyscanner's founder and getting considered for a job there. Yusuf then realised he could skip the recruiter and reach out directly to CXOs and share how he can add value to the firm, a process through which he met his Panalyt co-founder Daniel West, who was the CHRO International for Uber back in 2016.
Yusuf later joined Daniel on a consulting project and Dubai and got involved in the HR tech space. With a growing trend of niche HR tools, they identified a challenge for companies to extract actionable insights on the employee lifecycle from scattered employee data across various systems. This inspired them to start Panalyt, an HR analytics startup that provides HR Business Partners and People leaders with access to timely, relevant and actionable people data and insights, empowering them to make data-informed talent decisions.
Regarding Panalyt's expansion into Japan, Yusuf explains that they initially targeted Southeast Asia and India but faced challenges due to low willingness to pay and strong competition. However, an introduction to their co-founders in Japan helped them form a strong team with connections in the Japanese HR space.
They discovered that Japan lacked competition in the HR tech space, making it an ideal market. Additionally, the high labor costs and shrinking workforce in Japan created a demand for automation solutions, which Panalyt addressed successfully. Panalyt has since secured significant clients, including well-known Japanese corporations like Panasonic, Japan Tobacco International, and Sony.
Yusuf concludes by sharing a lesson he learned from his entrepreneurial journey: the importance of understanding market fit. He realized that the product-market fit played a significant role in their success, as they found the right market for Panalyt in Japan, where the demand for automation and HR tech was high.
Yusuf also explains the business applications of organizational network analytics (ONA) or social network analysis (SNA) in understanding communication patterns and behaviors within an organization. ONA involves applying graph theory to study the social connections between employees. It can be done through passive communication data from emails and chat metadata or through survey-based methods.
As Panalyt had already built an HR data warehouse, and they can leverage this data to create the social graph of a company. This allows them to explore interesting use cases, such as understanding communication patterns and their impact on performance and employee turnover risk.
The conversation then shifts to the focus of their product in Japan, which is more geared towards democratizing access to employee data and commonly requested talent-related metrics. Companies typically track numerous metrics on their employees, and various stakeholders, such as HR teams, managers, and CXOs, require specific data relevant to their roles. Panalyt automates this process, ensuring that each stakeholder gets access to the relevant data without manual intervention.
The decision to move to Japan was influenced by a recent regulation that required HR data disclosure, making their product highly relevant and valuable in the market.
Yusuf acknowledges that focusing on advanced analytics might not have been the best strategy for their market, but they saw interest from more mature markets like the US and Europe. However, they were not actively selling in those regions.
Yusuf concludes with some lessons learned, emphasizing the importance of understanding the market fit and tailoring the product's approach accordingly. He acknowledges that introducing advanced analytics to less mature markets can be overwhelming, and it's crucial to consider the market's readiness and receptivity to such solutions.
Links:
Enkrypt AI official website: https://www.enkryptai.com/
Sahil Agarwal LinkedIn: https://www.linkedin.com/in/sahil-agarwal/
Sahil Agarwal email: sahil@enkryptai.com
Panalyt official website: https://www.panalyt.com/
Yusuf Raza LinkedIn: https://www.linkedin.com/in/yusurfaza/
Michael Hutson LinkedIn: https://www.linkedin.com/in/hutsonmichael/
Please reach out to Yusuf Raza or Michael Hutson on LinkedIn if:
you are an entrepreneur with ties to Oxford and want to create more awareness about your venture
you are an investor interested in investing in Oxford linked startups
if you are a brand that wants to sponsor an episode of the OxIgnite podcast.
If you liked this episode, please help us spread the word to your network!
We would also welcome any donations to help us scale this podcast. You can donate at https://www.buymeacoffee.com/oxignite
Thanks for reading OxIgnite! Subscribe for free to receive new posts and support my work.