From cyber defence to care and contracts: AI100's eighth cohort puts AI to work
AI100's eighth cohort spotlights three leaders embedding intelligence into the systems enterprises rely on — data security, healthcare, and contracts.
The most useful artificial intelligence may not always announce itself with a chatbot or a flashy new interface. Sometimes, it is working quietly in the background — scanning a backup for an anomaly, bringing a patient's scattered history into focus, or turning a dense contract into a set of obligations a business can actually act on.
That quiet shift, from AI as an exciting add-on to AI as dependable enterprise infrastructure, is at the heart of AI100's eighth cohort.
AI100, an AWS and YourStory initiative, celebrates the pioneers, leaders, innovators, and evangelists shaping India's AI story. Across its cohorts, the series has followed technologists turning fast-moving research into tools that solve stubborn, real-world problems.
The latest lineup brings together three system builders working in very different domains. There is a cloud-native security leader making enterprise-grade data protection easier to use; a healthcare technologist proving that clean, connected data must come before intelligent care; and an enterprise software founder teaching contracts to do more than sit in a repository.
What connects them is not a fascination with AI for its own sake. It is an obsession with the foundations that make AI useful: sound architecture, trusted context, precision, explainability, and human judgment. Here is a peek at the three leaders making intelligence feel less like a layer and more like part of the plumbing.
David Gildea
David Gildea's journey towards AI began with a question he heard repeatedly as a consultant: "Can I just get this data in Excel?" Behind the request was a larger problem. Organizations had enormous volumes of information, but accessing, interpreting, and securing it remained painfully complex.
Gildea's early grounding was in data security and cryptography. Fresh out of college, his first publication explored how medical information could be stored securely on credit cards and made available to doctors when needed. Later, a startup immersed him in AWS and cloud-native design, where he saw how API-driven platforms could remove infrastructure friction and speed up problem-solving. That thinking led him to found CloudRanger, a SaaS company protecting customer data in cloud environments, which was eventually acquired by Druva.
Today, as VP of Product for Generative AI at Druva, Gildea leads the labs team looking two to three years ahead. His aim is deceptively simple: hide enormous security complexity behind plain-language, objective-driven experiences. Compliance teams can query systems for ISO or NIST reviews, while AI agents surface gaps, estimate timelines and costs, and outline remediation. Druva also analyzes backup data across workloads to identify anomalies and support investigations, while its Managed Detection and Response service flags suspicious patterns before incidents escalate. DruAI now resolves 68% of customer issues directly.
For Gildea, however, speed cannot come at the cost of safety. Druva has delayed full MCP deployment while security specifications mature - a reminder that, in cyber defence, caution is part of the product. Transparency matters just as much. "Sunlight is the best cleanser," he says. His larger mission is to democratize data security without diluting the enterprise-grade protection beneath it. Away from the labs, he resets by golfing with his son, spending time with family, and heading outdoors to beaches and hiking trails.
Ankit Maheshwari
If Gildea is stripping complexity out of security, Ankit Maheshwari is tackling one of healthcare's oldest technology problems: intelligence cannot travel through data that does not connect.
Maheshwari, Chief Product and Technology Officer and founding team member at Innovaccer, remembers the first time a few lines of code created something real. By his second year of engineering, he was working in his college's software incubator, experimenting with an early lecture-transcription and indexing tool as well as a camera-based attendance system. Those projects shaped a belief that still guides him: intelligent systems should amplify human capability, not replace it.
Innovaccer began as a horizontal data and analytics company, helping researchers at institutions including Stanford, Harvard, Wharton, and MIT work with analytics-ready datasets. After projects across sectors, including NASA and Walt Disney, the team encountered a healthcare customer in Des Moines, Iowa. The engagement exposed a vast contradiction: the US spent nearly $4 trillion a year on healthcare, yet siloed information still kept patient histories and timely insights away from providers. Innovaccer made a decisive pivot to healthcare.
Maheshwari's architecture philosophy is infrastructure-first. Instead of bolting chatbots onto legacy systems, Innovaccer starts with data unification, analytics, and longitudinal patient context, then builds actionable intelligence for care teams. The platform now supports more than 80 million lives and powers seven of the top 10 health systems in the US.
The hard part is not only technical. Healthcare comes with strict compliance, decades-old systems, and clinicians whose workflows cannot be disrupted for novelty's sake. New tools must earn trust and fit naturally into the working day. Looking ahead, Maheshwari is watching edge AI for faster, more secure on-device intelligence and robotics for AI that can operate in the physical world. But the near-term lesson remains grounded: clean data, reliable infrastructure, and human-centred design are what turn AI ambition into better care. Off the clock, cricket, non-fiction, and poker give him three very different ways to think about performance, leadership, probability, and people.
Aditya Gupta
For Aditya Gupta, the next enterprise interface may be hiding in a document most people open reluctantly: the contract.
The Co-founder and Chief Technology Officer of Sirion wants to transform contracts from static files into intelligent systems that can understand context, surface risk, track obligations, and participate in business workflows. His approach has roots in an unlikely place - autonomous robots designed for Mars exploration.
At IIT Kharagpur, Gupta worked on systems that had to function with limited human intervention, communication gaps, and uncertain conditions. The experience taught him to decompose ambiguity into manageable parts and design for reliability - lessons that now inform technology used at enterprise scale.
Founded in 2012 by Gupta, Ajay Agarwal, and Claude Marais, Sirion builds AI-powered contract lifecycle management software. The founders saw that contracts govern critical relationships and outcomes, yet enterprises often treated them as rigid documents scattered across repositories. Sirion instead breaks contracts into actionable components so obligations can be monitored, risks analyzed, and decisions informed by historical context.
AI is not a layer on top of that system. It sits at the core, supporting clause extraction, risk analysis, obligation tracking, drafting, negotiation, and post-signature management. As the platform moves towards agentic systems, AI can guide next steps and, in high-confidence scenarios, take action within a workflow.
In legal technology, though, "almost correct" is not good enough. Sirion grounds its models with a legal knowledge graph and ontology, pairs recommendations with references, and designs for explainability, auditability, and scale. Gupta describes AI decisions as probabilistic: moving towards 98-99% confidence helps, but trust ultimately comes from grounding an answer and showing users why it is reliable. His principle is clear: "AI should augment human judgment, not replace it." Daily reading keeps his technical curiosity moving; conversations with his children often supply the unexpected questions that open a new perspective.
Across Druva, Innovaccer, and Sirion, AI is becoming most powerful when it recedes from view. The interface gets simpler, while the system underneath grows more rigorous. Security needs transparency. Healthcare needs connected data and workflow trust. Contracts need domain context and explainable decisions.
That is the common thread running through AI100's eighth cohort: before AI can act intelligently, it must be engineered responsibly. Gildea, Maheshwari, and Gupta are not simply adding AI to enterprise software. They are rebuilding the foundations so intelligence can work where the stakes are highest.