Rajeev Sharma, CTO at Grid Dynamics — GenAI, Continuous Innovation, Rocket Propulsion Challenges, Leadership, Skills for Engineering Leaders, Ethical AI, Business Reimagination

In the dynamic world of technology, staying ahead of the curve is not just an advantage—it is a necessity. Rajeev Sharma, the Chief Technology Officer at Grid Dynamics, exemplifies this ethos through his leadership and innovative vision. From developing rocket propulsion systems to spearheading AI-driven business solutions, Sharma’s journey is a testament to the power of continuous innovation. In this exclusive interview, he shares his strategies for fostering a culture of creativity, the impact of Grid Lab, and the transformative potential of AI in reshaping business models for Fortune 1000 companies. Join us as we explore the future of technology through the eyes of a pioneer.

As the CTO of Grid Dynamics, how do you foster a culture of continuous innovation and keep your team motivated amidst rapid technological advancements?

At its core, Grid Dynamics embraces the philosophy of “Doing the right thing for the customer”. Our capacity for delivering high-quality work stems from our deep-rooted culture of engineering rigor, a collaborative approach powered by our globally distributed teams, a drive for innovation, and a penchant for pushing the envelope when it comes to solving complex problems. Our engineers are a talented and confident collective, embodying a principled approach to data-driven decision-making. They articulate their engineering insights with conviction, presenting solutions meticulously underpinned by rigorous data analysis and structured hypotheses.

The Grid Lab, our internal R&D hub, serves as a fountainhead of new ideas and innovative technical solutions, and acts as a training ground for our engineers. In this lab, we work on projects that are inspired by the challenges that our customers face. We invest millions of dollars and run these projects in mission mode with a strict eye on business value for our customers and us.

Internally, we provide numerous platforms for collaboration and knowledge sharing, such as Dynamics Talks, Architecture forums, and bi-weekly and monthly interactions covering emerging technology paradigms. In these forums, our engineers and architects present their ideas and receive instant, constructive feedback from the large pool of engineers who regularly attend these sessions.

In the past year alone, we have completed more than 30 POCs and demos on generative AI (GenAI), spanning all three hyperscalers (Microsoft, Amazon and Google cloud providers). The use cases span multiple industries, and many POCs are now ripe for scaling up.

We foster a strong culture of continuous innovation by offering a plethora of opportunities for our engineers to reskill and upskill, as well as engaging them in extremely challenging projects for Fortune 1000 companies. The CTO office is in the thick of all the engineering and technology fervor driving these interactions.

Your work on the Design of Agni-III Solid Rocket Propulsion systems earned you the Scientist of the Year award. Can you share some key challenges you faced during this project and how you overcame them?

As a young and newly minted Army Major, designing Stage-1 and Stage-2 solid rocket motor cases and the retro-rocket for assisting stage separation was indeed an honor and a privilege. Looking back, I must say that the challenges that unfolded were a captivating amalgamation of technical complexities and the nuanced, yet potent, human elements intrinsic to any large R&D institution involving multiple stakeholders.

From a technology standpoint, the challenges were manifold, as the Defence Research and Development Organisation (DRDO) had no precedent for designing rocket motor cases of such imposing diameter using ultra-high-strength high-alloy steel. Fabrication intricacies, machine tolerances, tooling and fixture design, and the judicious selection of design criteria for safety factors—striking a delicate balance between excessive weight, which is a big ‘No’ for all flight-worthy systems, and not enough weight, which could jeopardize the entire mission—presented formidable challenges. Stress analysis of the bolted joints, the clevis-tang joint (yes, this was the very same joint configuration that led to the Challenger mission failure), and the flex nozzle for thrust vector control further compounded the technical hurdles.

During those days, I worked relentlessly to overcome multiple conflicting design imperatives, managing an intricate web of stakeholders spanning aero-structures, aerodynamics, onboard computer systems, propulsion systems, and beyond. The culmination of these efforts was the successful pressure testing, including burst testing, of the very first set of Stage-1 and Stage-2 systems, paving the way for further ground and flight tests. In hindsight, the journey was a narrative of success, but it was indeed a tumultuous odyssey, especially for a young rocket scientist working under immense pressure to succeed in the very first iteration. I am glad it all turned out well.

How do you see the role of artificial intelligence evolving in the next decade, particularly in relation to business automation?

This is a hard one to answer given the pace at which developments in AI, and especially GenAI, are taking place. However, it is prudent to anticipate an infusion of AI/GenAI-powered capabilities permeating the underlying business processes of all applications, be they B2B, B2C, B2B2C, or P2P. This pervasive integration will foster a gradual yet inexorable shift, as the human race grows increasingly comfortable with the ubiquitous presence of AI-driven operations.

However, a pivotal challenge that demands our collective attention lies in cultivating trust in the outcomes generated by these AI systems. This endeavor necessitates a multi-disciplinary approach, one that rigorously addresses the issues of trust and ethics in AI-based systems, particularly in highly regulated industries such as healthcare, financial services, and processes involving PII or business-critical data.

The advent of GenAI has also reignited our focus on UX and natural-language-based conversational systems (like chatbots), which can serve as gateways to orchestrate a symphony of multi-agents and/or multi-modal—Large Language Models (LLMs) and Large Vision Models (LVMs)—operations like coding, product design, legacy code modernization, etc. This paradigm shift will undoubtedly spawn new specializations, akin to the emerging role of “Prompt Engineers.”

Moreover, we are likely to see an increasing number of use cases that transcend the boundaries of individual models, seamlessly integrating deep learning, machine learning, and other paradigms. These models will be invoked via LLMs connected to domain-specific document corpora, employing techniques such as Retrieval-Augmented Generation (RAG), with results elegantly communicated through chatbots, voice interfaces, or visually compelling dashboards and graphical plots.

Notably, the inevitable confluence of AI, cybersecurity and quantum computing in the not-too-distant future also promises to reshape the technological landscape in profound ways.

Last but not least, we can anticipate the emergence of new business operating models equivalent to the disruptive forces of SaaS, DaaS, and IaaS, ushering in a new vanguard of winners and leaders on the block. As compute and storage pressures mount, it will be interesting to see how the hyperscalers and SaaS solutions perform against the emerging story of GenAI and its impact on developer productivity and digital engineering as a whole.

In conclusion, humanity will be at the center of technological advancement, where intuitive user experiences and very tightly coupled human-computer interaction (HCI) will become a default.

What are some common misconceptions about AI and automation in the business world, and how do you address these when discussing potential AI solutions with stakeholders?

There are many misplaced beliefs and common misconceptions about AI. These misguided notions range from the reductive belief that AI is solely about automation to the existential dread of intelligent machines displacing human workers. This fear includes the idea of rendering entire professions like coding, finance, accounting, legal, and back-office operations obsolete, and upending the socio-economic fabric. Additionally, there are unfounded apprehensions about rogue AI systems taking over the planet, and the inherent untrustworthiness of AI outputs due to perceived biases.

To dispel these misconceptions, a multi-pronged approach is essential:

• Foster a constant dialogue and implement organization-wide training initiatives to raise awareness and promote a deeper understanding of AI’s capabilities and limitations.
• Provide opportunities for reskilling and upskilling, empowering the workforce to adapt and thrive in an AI-driven landscape.
• Strategically distribute AI-savvy and digital-savvy talent across all functional groups within the enterprise, ensuring a pervasive integration of AI capabilities.
• Champion diversity and inclusion, and empower the workforce with the right tools, training opportunities, active coaching, and mentorship.
• Help the workforce understand the fundamentals behind model training, data sources, and the checks and balances/guardrails employed to ensure safe, ethical, and trustworthy outcomes.

By addressing these misconceptions head-on, we can cultivate a culture of curiosity, innovation, and collaboration. In such a culture, AI is embraced as a powerful tool to augment human capabilities rather than a threat to job security, societal stability, or the relevance of specific professions. Through continuous education, skill development, and a commitment to ethical AI practices, we can harness the transformative potential of this technology while mitigating its risks and addressing legitimate concerns.

How has your academic background in Management & Systems Design from MIT Sloan and Space Engineering & Rocketry from BIT MESRA influenced your leadership style and strategic decision-making at Grid Dynamics?

From my formative days donning military fatigues, spearheading cutting-edge innovations in rocket propulsion systems, to later adorning a corporate suit and tie, continuously shaping the narrative of how technologies catalyze business value creation, my leadership style has evolved over time. While the foundational tenets of leadership—loyalty, integrity, honesty, and a high order of professional competence—remain immutable across domains, I have learned to adapt my style to empower our knowledge workers to thrive. In stark contrast to the military, where directives are followed with unwavering obedience, the knowledge workforce thrives in an environment that fosters intellectual freedom, embraces iterative learning, and treats failures as opportunities for growth.

The intensely interdisciplinary nature of the aerospace industry instilled in me a deep reverence for systems thinking, systems design, non-linear thinking, and the ability to solve complex problems against very tight deadlines and mission constraints. My time at MIT enriched this perspective, exposing me to many different areas such as real options in large complex systems design and development, multidisciplinary-systems design optimization, foundations of strong product design, and systems engineering.

Throughout my professional and very intense academic journey, the heuristics and frameworks for problem-solving I have cultivated have held me in good stead. Looking back, all these institutions and the leaders therein have shaped me to become the professional and the human being I am today.

Can you provide an example of a recent AI-driven project at Grid Dynamics that has had a substantial impact on your clients’ business operations?

At Grid Dynamics, we have been at the vanguard of harnessing the transformative potential of artificial intelligence to drive substantial impact on our clients’ business operations. However, AI-based projects don’t happen in a vacuum. There is a requisite level of digital savviness and readiness that must be nurtured at all echelons of a large enterprise before they have the muscle to successfully infuse machine intelligence into their business operating model. We have done some amazing work over the past 18 months in the areas of cloud, data and AI engineering, spanning both deep learning and machine learning use cases, as well as those powered by the more recent advancements in GenAI.

Without going into specific proprietary details, a few examples of the many AI-based enterprise solutions we have built are provided below:

• We built a price optimization engine that drives targeted promotions for a major grocery store chain, leveraging AI to enhance their pricing strategies and customer engagement.
• We developed a GenAI-based conversational assistant tailored for financial advisers in the wealth management and financial services sector, streamlining their operations and enhancing client interactions.
• We have proven the efficacy of LLMs for legacy code migration, enabling the seamless transition from legacy technologies like RPG and Cobol to modern, high-level technologies such as Java. This has been instrumental in our UI replatforming projects for an automotive customer, facilitating the conversion of code from REACT to Next.js.
• We developed solutions powered by vision models and LLMs to accelerate product design processes, reducing the time required to convert 2D engineering drawings into 3D renderings—a capability that has proven extremely useful for our manufacturing customers, enabling them to streamline their product development lifecycles.
• We implemented a GenAI-powered product data enrichment solution for a major retailer to generate compelling, personalized, multilingual product titles, descriptions, attributes and SEO metadata, accelerating product onboarding and enhancing customer experience.
• We are developing a GenAI virtual try-on and product visualization and customization solution for a global apparel brand to enhance the online shopping experience and boost customer engagement.
• We are one of the leading AI services companies specializing in multi-agent, multi-modal (LLMs and LVMs) models for various use cases in alternative investments within the finance sector, particularly in wealth management. All of our POCs in this area require advanced RAG techniques, along with fine-tuning methodologies and architectural decisions related to vector databases and semantic caching.

Underpinning all the above AI and GenAI solutions is our deep expertise spanning more than 8 years in AI, cloud and data engineering, coupled with our strong experience in UX design for building innovative products and platforms.

In your opinion, what are the most critical skills that engineering leaders need to develop to effectively manage the intersection of AI and business?

The advances in AI, and particularly GenAI, are taking place at such a breakneck speed that it’s almost impossible to imagine any application being built without harnessing the power of an underlying AI engine(s). The infusion of machine intelligence into a business operating model necessitates constructing a comprehensive digital fabric that permeates every layer of the technology foundation ecosystem—infrastructure, data, business processes, the front-end layer, and the glue of a well-designed API ecosystem—bringing the whole digital continuum to life.

The enterprise architecture of today’s digitally powered business is a journey of “System of Systems”, characterized by socio-technical systems, loosely coupled business processes encapsulated in the notion of a microservices archetype, and a well-oiled, highly automated environment powered by continuous integration-continuous delivery (CI-CD) processes. Managing such a complex, transient, and responsive web of technologies in any large-scale enterprise imposes immense pressure on its leaders. The attributes in the following indicative, yet not exhaustive list, are critical enablers of success:

• Leaders must understand the principles of systems design, systems thinking, and enterprise architecture, and how these elements integrate into the broader vision of the company and its position in the industry and markets.
• A digitally savvy mindset and a sophisticated grasp of API-led digital enterprise design principles and the importance of a scalable and tunable infrastructure (read software-defined infrastructure) are essential.
• Effective navigation of organizational complexities requires assertive and convincing leadership (read soft skills), particularly in dealing with organizational silos. Leaders need to be comfortable with “reimagination” and “reinvention” as operative terms for pushing the boundaries of competitive advantage in a hyper-connected, AI-infused society and enterprises.
• Leaders need to be high on the AI-savviness index, with a clear comprehension of what AI can and cannot do, and how to drive its adoption across different business processes to achieve optimal and tangible business impact.
• Leaders must champion the AI adoption agenda by building a cross-functional team of leaders encompassing all functional units of the enterprise

What are some of the ethical considerations you take into account when implementing AI and automation solutions, and how do you ensure these are addressed?

When considering infusing AI and machine intelligence into the business operating model of an enterprise, establishing trust in the outputs of the AI model is critical. This necessitates unambiguously defining the boundaries of acceptability, and continuously employing appropriate metrics to monitor and mitigate bias in the AI model’s output. Quite honestly, however, this is an inherently complex path to traverse, and the real value lies in the effective implementation of such an approach. At a more granular level, some suggested methods for ensuring ethical AI implementation include:

• Effectively implementing guardrails to prevent unauthorized access, and ethics-based checks as output scanners;
• Thoroughly examining underlying documentation for potential conflicts or ethical concerns before implementing a RAG solution (in the case of GenAI-based solutions);
• Ensuring realistic sampling of data sources to mitigate bias and promote representative outcomes;
• Providing citations and retrieval statistics, along with document classification from retrieval processes to promote transparency and accountability;
• Aligning open source models toward safer responses using LoRA (Low-Rank Adaptation) techniques;
• Proactively creating refusal scenarios while developing applications to establish ethical boundaries; and
• Ensuring the publication of model and data scorecards for each project, allowing for a quick overview of the model’s capabilities and performance across cohorts.

While these measures are necessary, they are not sufficient in isolation. Building an ethical, trustworthy and intelligent AI platform is a collaborative endeavor, requiring the harmonious convergence of various functional units within the enterprise, tailored to its unique context and industry. No silver bullet exists here; rather, a multidisciplinary approach is essential to navigate the ethical complexities inherent in AI adoption.

How do you foresee AI and automation transforming traditional business models, and what advice would you give to companies looking to stay competitive in this evolving landscape?

To succeed and thrive in the digital economy, the imperative for leadership is the creation of mechanisms that foster end-to-end business reimagination. The paradigm of “if it ain’t broke, don’t fix it” has lost its relevance in the face of rapid technological advancement. The confluence of data in all formats, scalable cloud architectures, and the disruptive potential of AI and now GenAI compels every business, across every industry, to reimagine and recreate its processes through intense scrutiny. The unified goal? Building core competencies & deepening the barriers to entry for competitors.

Investments in building scalable, intelligent (AI-powered) digital platforms in every industry can uncover new business opportunities like never before. The effective use of the trifecta—data, cloud, and AI—can now change the narrative of growth and profitability. However, the capacity to build a strong digital business foundation goes way beyond a few isolated pilots. It’s a team game between business, technology and human experience designers, who must intelligently craft a business playbook that is hard for the competition to replicate in the short and medium term.

The ability to serve the market with a high dose of machine-augmented intelligence and unleash autonomous business actions is the key to long-term market domination. As ever, the propulsive power of leadership will matter the most—it’s an opportunity cost for every enterprise.