David Gildia's career has been shaped by the constant question from clients: 'Can I just get this data in Excel?' As a consultant, he personally observed how organizations struggled with accessing, interpreting, and protecting their own information. The problem was not only technical but also human. Realizing that finding patterns in massive datasets exceeded human capabilities, Gildia turned his attention to artificial intelligence (AI) as a way to make data both usable and secure at scale.
However, his initial specialization was in cryptography and data security. Immediately after university, his first publication concerned enabling the secure storage of medical information on credit cards, allowing doctors worldwide to access personal medical data when necessary. Although machine learning (ML) and AI played a limited role in testing and verification, they were not central to his work. Only years later did Gildia begin to view AI as an essential tool for solving real-world security challenges.
He notes: 'AI is almost magical. It has the ability to analyze huge volumes of data and find what is inaccessible to humans.' Gildia's transition from cloud technologies to AI was a natural evolution. Working at a startup immersed him in AWS technologies and cloud-native design, and he quickly embraced the flexibility based on cloud APIs. In his view, traditional data centers became superfluous, while cloud platforms allowed teams to test, iterate, and solve problems faster. This approach prompted him to found CloudRanger—a SaaS company specializing in securing customer data in the cloud environment. Its cloud approach aligned with Druva's global mission of data protection, ultimately leading to its acquisition.
Over time, Gildia began to see strong parallels between the growth of cloud computing and the adoption of generative AI. He asserts: 'The journey to the cloud is very similar to the journey in AI. The cloud simplified client problem-solving, and AI does the same, perhaps even more effectively.' Unlike cloud infrastructure, which still requires technical expertise, AI has an exceptionally low barrier to entry. 'Even at home, everyone uses AI 10-15 times a day. There is no barrier.'
Today, serving as Vice President of Product for Generative AI at Druva, Gildia leads a laboratory team focused on how products will evolve over the next two to three years. The goal is ambitious: to expand the reach of enterprise-grade data protection and make it accessible to organizations using GenAI. At Druva, generative AI allows users to interact with complex security systems in plain language. For example, compliance teams can directly query Druva for reviews against ISO or NIST standards. AI agents then generate detailed reports outlining gaps, assessing timelines and costs, and describing troubleshooting steps—all without using spreadsheets or complex tooling. Gildia emphasizes: 'GenAI allows us to hide immense complexity behind a simple, goal-oriented interface. Our customers gain powerful capabilities without needing deep technical expertise.'
Druva's AI-driven approach also strengthens the company's core security operations. The company manages thousands of daily backups for workloads such as VMware, EC2, endpoints, Azure, and cloud environments. By analyzing large volumes of backup data, AI identifies significant patterns and anomalies, supporting cyber threat investigations and compliance checks. Users can query the system using frameworks like NIST or MITRE and quickly generate board-ready reports detailing timelines, fixes, and impact. The recently launched MDDR (Managed Detection and Response) service goes further by proactively flagging anomalies in backup data, often detecting issues before incidents escalate. 'We have used AI for years to improve the customer experience. Generative AI is critical because it allows us to democratize data security across all organizations, instead of limiting it to those working in the backend.'
One of the most significant challenges Gildia faces is the incredible pace of AI development combined with rising customer expectations. The modern customer has practical experience using tools like ChatGPT and Gemini and expects the same level of speed, accuracy, and clarity from enterprise software. In response, Druva's tools and services are engineered for high precision, handling thousands of customer requests, both large and small, with 68% of customer issues now resolved directly through DruAI. Nevertheless, security remains non-negotiable. At this early and rapidly changing stage of AI, Gildia believes that cutting corners can quickly become a source of problems. Therefore, Druva maintains a security-first approach, designing AI systems with the same enterprise-grade protection guarantees as its core products. This means prioritizing caution over speed. For instance, the company delayed the full rollout of MCP because security specifications are still being refined, despite achievements like AWS Age Gateways. While this approach may slow down releases, it ensures vulnerabilities do not reach customers, turning security into a long-term competitive advantage.
For Gildia, transparency is the cornerstone of AI ethics. He is convinced that companies must openly disclose how they create, train, and deploy AI systems, creating space for scrutiny and accountability. He states: 'Light is the best disinfectant,' underscoring the importance of openness in responsible AI development. Gildia also views AI as both a powerful opportunity and a growing threat. He urges organizations to adopt it as a line of defense. AI-driven threats are becoming increasingly sophisticated, and he points to emerging malware capable of rewriting its own code during execution to evade antivirus detection. As attacks grow, AI-based countermeasures will become indispensable.
At Druva, AI is embedded as a default layer of defense to help customers protect their environments. Gildia notes that data security is a constant cat-and-mouse game requiring immediate action and investment. The best time to act is now, before threats become even harder to detect and contain. Druva's long-term partnership with AWS is based on the founders' early conviction that the future of enterprise software lies in SaaS and cloud services. Recognizing AWS as the leader in the cloud market, Druva initially built its work on this platform, integrating new services like DynamoDB, S3, EC2, and RDS as they launched for customer benefit. Today, Druva utilizes 65-70% of AWS services globally, with a presence in most data centers. This collaboration is deeply symbiotic, merging AWS's global, problem-solving approach with Druva's focus on data protection and security. Druva provides early feedback, collaborates with AWS product managers and engineers, and actively participates in AI-related product development. This includes joint initiatives in the AWS Gen AI Innovation Center, where teams co-create MVPs through coding sessions. Gildia remarks: 'This is a great way to work with Amazon—line by line in the code.'
Gildia advises IT leaders to embrace the potential of AI, especially in strengthening data protection. He believes organizations should bring together employees from different functions to understand how AI can be leveraged, what tools are available, and what risks and opportunities they present. This is key to transitioning from a traditional user experience to a more conversational, AI-managed interaction. To stay ahead in this fast-moving field, Gildia reads widely, attends industry conferences such as AWS re:Invent, rapidly builds Minimum Viable Products (MVPs) using tools like AWS CodeWhisperer, and follows leading voices on LinkedIn. Outside of work, he relaxes by playing golf with his son, spending time with family, visiting beaches, and hiking to enjoy nature.
The most useful artificial intelligence does not always attract attention through chatbots or flashy interfaces. Often, it works unnoticed in the background: scanning backups for anomalies, focusing scattered patient histories, or transforming complex contracts into a set of obligations that business can work with.
This quiet shift from AI as an exciting add-on to AI as reliable enterprise infrastructure is at the heart of the eighth cohort of AI100. This initiative, a joint project by AWS and YourStory, recognizes innovators, leaders, developers, and evangelists shaping the history of AI in India. As part of its series, the program tracks technology specialists who are turning rapidly evolving research into tools that solve complex real-world problems.
The latest group comprises three system developers working in completely different fields. Among them are a cloud security lead simplifying the use of enterprise-level data protection; a healthcare technologist demonstrating that clean, connected data must precede intelligent care; and a founder of an enterprise software development company who teaches contracts to do more than just reside in a repository. What unites them is not just an interest in AI, but an obsession with the fundamental aspects that make AI useful: robust architecture, trusted context, accuracy, explainability, and human judgment. These leaders make intelligence less like an add-on layer and more like an integral part of the system.
Gildea's journey into AI began with a question he often heard as a consultant: 'Can I just get this data in Excel?' Behind this query was a broader problem: organizations possessed vast amounts of information, but access to it, its interpretation, and its protection remained extremely difficult. His early experience focused on data security and cryptography. Immediately after graduation, his first publication explored how medical information could be securely stored on credit cards and made available to doctors when necessary.
Later, the startup immersed him in AWS and cloud-native design, where he saw how API-managed platforms could eliminate infrastructural hurdles and accelerate problem-solving. This experience led to the creation of CloudRanger, a SaaS company protecting client data in cloud environments, which was subsequently acquired by Druva. Today, serving as Vice President of Product for Generative AI at Druva, Gildea leads a laboratory team focused on a two-to-three-year horizon. His goal seems simple: to hide immense security complexity behind simple, goal-oriented user scenarios.
Compliance teams can request systems for checks against ISO or NIST standards, while AI agents identify gaps, assess timelines and costs, and describe remediation paths. Druva also analyzes data backups across various workloads to detect anomalies and support investigations, and its Managed Detection and Response service flags suspicious patterns before incident escalation. The DruAI service now solves 68% of customer issues directly. However, for Gildea, speed should not come at the expense of security. Druva has postponed full MCP implementation until security specifications mature—a reminder that caution is part of the product in cybersecurity. He emphasizes that transparency is equally important: 'Sunlight is the best disinfectant.' His broader mission is to democratize data security without compromising enterprise-grade protection.
In the modern corporate landscape, contracts play an omnipresent role, regulating relationships, defining obligations, and shaping business outcomes. However, for decades, they have remained static, opaque, and difficult to understand.
Aditya Gupta aims to change this situation. As Co-founder and CTO of Sirion, he is leading the transition from contracts as passive documents to contracts as intelligent systems powered by artificial intelligence, capable of reasoning, and integrated directly into business workflows.
Yet, his journey to this point began far from enterprise software, rooted in a deep interest in how systems function beneath the surface. Gupta started programming early, drawn to the idea that what is created once can be scaled infinitely. Over time, this interest transformed from writing code to designing systems capable of serving thousands, and then millions, of users. This progression led him to enterprise software, and subsequently to artificial intelligence, where systems move beyond deterministic logic to support reasoning and decision-making.
At the Indian Institute of Technology (IIT) in Kharagpur, Gupta immersed himself in an environment that emphasized first-principles thinking, problem-solving, and peer learning. Alongside his classmates, he learned to break down complex, ambiguous problems into solvable components.
One significant experience was working on autonomous robotic systems designed for Mars exploration. In such scenarios, human intervention is limited, forcing teams to consider communication gaps, system failures, and machine collaboration. The task was not just engineering, but developing systems capable of operating reliably in uncertain conditions.
By breaking these problems into smaller, manageable parts, Gupta and his team managed the complexity and created functional solutions. These early experiences continue to influence his approach today, combining systems design, scalability, and intelligent automation.
Sirion, founded in 2012 by Gupta, Ajay Agrawal, and Claude Marey, offers AI-powered contract lifecycle management software. Contract management, covering all stages from creation to renewal, has traditionally been a fragmented and error-prone process, despite its central role in business operations.
The founders of Sirion identified a critical gap: contracts form the backbone of enterprises, yet they are treated like rigid, static documents. Users often struggle to locate agreements, interpret obligations, or assess whether new contracts conflict with existing terms. While many CLM tools focused on storage and workflows, they largely ignored intelligence and usability.
Sirion decided to change this approach. Instead of viewing contracts as files stored in repositories, the company began to see them as living systems—entities that must be continuously understood, tracked, and acted upon. By combining Gupta's technical expertise with the legal experience of his co-founders, Sirion positioned itself at the intersection of law and technology.
A key part of this transformation was segmenting contracts into smaller, actionable components. Obligations could then be tracked, monitored, and reported in real-time, ensuring that decisions were based on historical context rather than being created from scratch.
Gupta notes: 'My role has evolved over time, but it is generally at the intersection of AI architecture, product, and strategy. One conscious choice I made was that AI would not be just an add-on layer. At Sirion, it is embedded into the core system. Whether it is clause extraction, risk analysis, or obligation tracking, AI is deeply integrated.'
For Gupta, contract management, like the underlying technology, must remain flexible and adaptive. The Sirion platform utilizes AI and Generative AI (GenAI) and is increasingly moving toward agentic systems. In this model, AI does more than just analyze or generate; it actively participates in workflows.
Sirion's systems are designed to understand contracts end-to-end: analyzing documents, identifying risks, and guiding users on next steps. By leveraging historical data and contextual information, AI surfaces relevant insights directly, reducing the need to manually sift through lengthy documents. In high-confidence scenarios, the system can even act autonomously, accelerating the decision-making process.
He emphasizes that there has also been a shift in the application of generative AI across the entire contract lifecycle. From drafting and negotiation to post-signature management, AI maintains context across different stages, transforming contracts into dynamic intelligence systems. These systems do not merely store information; they actively inform decisions, manage relationships, and drive outcomes.
Nevertheless, implementing AI in the legal domain comes with unique challenges. Accuracy is a non-negotiable requirement: 'almost right' is insufficient. Simultaneously, users must understand how and why decisions are made, making explainability critical. Scale adds another layer of complexity, as enterprises manage millions of contracts in various formats. Underlying it all is the issue of trust.
To address these issues, Sirion created a legal knowledge graph and ontology that anchors AI outputs in domain-specific understanding. Every recommendation is accompanied by clear citations, ensuring transparency. System design focuses on scalability and fault tolerance to handle enterprise-level workloads.
For Gupta, trust arises at the intersection of accuracy and clarity. By making AI decisions transparent and context-driven, Sirion enables users to rely confidently on intelligent systems.
He acknowledges that AI is still evolving. 'Nothing in AI is 100%. The closer we get to the high 90s, 98–99%, the more confident users become. Every decision remains probabilistic, but the system's role is to justify that probability and provide assurance that the outcome is reliable,' he explains.
Gupta views AI as a fundamental pillar of the future, comparable to the shift from telephony to the internet. Its applications span industries from healthcare and climate modeling to scientific discovery.
However, the real challenge lies not in creating possibilities, but in implementing them responsibly. Issues of governance, fairness, and equitable access remain central to the evolution of AI.
Gupta is particularly interested in the rise of agentic systems, which signal a shift from reactive tools to proactive systems capable of operating within workflows. He also points to the convergence of AI with knowledge graphs and the growing emphasis on domain-specific intelligence embedded in enterprise platforms.
He stresses that ethics is an integral part of this journey. Data bias can skew results, making meticulous data curation and the creation of 'golden records' crucial. Accountability and auditability ensure that every AI-driven decision can be traced.
'AI should augment human judgment, not replace it,' says Gupta, emphasizing the importance of transparency, fairness, and human oversight.
Beyond AI, he continues to monitor developments in space exploration, quantum computing, and emerging model architectures—areas that reflect his enduring interest in complex systems.
Sirion's technological path has been closely tied to AWS since 2014. Starting with core compute services, the company gradually expanded its stack to include serverless architectures and, more recently, AI capabilities through Bedrock. AWS has allowed Sirion to scale efficiently, offering a robust, globally deployed, and compliant infrastructure. The partnership extends beyond infrastructure, including close collaboration between Sirion and AWS's research and machine learning teams.
This joint implementation has helped Sirion tackle complex problems, refine model strategies, and accelerate its AI roadmap while maintaining stable performance and scalability.
For Gupta, maintaining the pace of AI development requires both discipline and curiosity. He dedicates time daily to reading about new research, technological shifts, and innovations in fields such as healthcare, generative AI, and quantum computing. Conversations are equally important, whether with colleagues from different disciplines or even with his children, whose questions often spark new perspectives. 'Inspiration is everywhere,' he says.
At the same time, he emphasizes the importance of balance. In a rapidly changing technological landscape, it is vital to stay grounded. Continuous learning, open dialogue, and time spent away from the screen help him maintain the clarity necessary for meaningful innovation.
From designing autonomous systems for Mars to creating intelligent contract ecosystems, Gupta's journey reflects a constant thread: curiosity transformed into scalable systems. At Sirion, this curiosity is now shaping the future of how businesses understand and act upon their most critical documents.
Russian specialists have developed a special unit capable of extracting up to a thousand liters of drinking water from atmospheric air daily. This technology was first successfully tested in the arid climate of Uzbekistan and subsequently in India.
The system has been named 'Water in the Desert' and is designed to operate across a wide temperature range, from plus twelve to plus sixty degrees Celsius. According to the creators, one basic set of equipment is sufficient to supply water to small settlements, tourist camps, industrial facilities, or remote field bases.
The tests in Uzbekistan were particularly significant because they were conducted in 2024 under conditions of severe drought. This allowed for testing the system's functionality in areas where the combination of high temperature and low humidity significantly complicates the process of extracting water from the air. In 2025, testing continued in India.
The principle of operation involves the unit drawing in surrounding atmospheric air and passing it through several preparatory stages. Subsequently, the airflow is directed to cooling plates. As a result of cooling, the water vapor contained in the air condenses and transforms into liquid water.
After this, the treated air enters a recuperator—a device designed to return some heat and transfer it to the next stream of air. The developers note that thanks to this scheme, energy consumption has been reduced to 0.4–0.7 kilowatt-hours per liter of water obtained.
A key advantage of this unit is its ability to operate effectively even at very low humidity levels, as low as five grams per cubic meter of air. This makes it suitable for use in the arid regions of Central Asia, as well as in the Middle East and Africa, where other systems fail. For comparison, devices based on direct cooling, like air conditioners, typically require at least eight to ten grams per cubic meter.
It was previously noted that Central Asia has faced an unprecedented shortage of water resources due to a five-year drought. The period from 2021 to 2025 has been recognized as the hottest for the region in the last 40 years, and soil moisture levels have reached historically low marks.
