I'm Rahul Gupta. Twenty-six years running technology at scale — first-ever CTO, then CTPO, of Conduent (NYSE: CNDT, ~$6B) and CEO, Technology Services at Toptal — and now building Kaia AI: self-improving, human-in-the-loop AI for the industries that can't afford to get it wrong.
Building and running technology businesses
Bespoke platforms consolidated into standard products
Technology-led revenue, grown from $800M
AI/ML products shipped from an Innovation Hub of 300+ engineers
After running an 8,000-person technology organization and a $400M services P&L, I write and ship again — because the fastest way to test a thesis about AI is to build with it.
Kaia AI is a self-improving, human-in-the-loop AI platform for regulated industries. It's built on a conviction earned over 26 years of operating: in regulated work, AI only scales when trust is engineered in — pretrained vertical AI agents, human-in-the-loop correction that makes the system better with every expert review, governance by design, and benchmarks you publish rather than promise.
Kaia is a founder-stage venture and a working proof of the thesis, not a victory lap. Every correction a domain expert makes, makes the platform smarter — that is what self-improving means.
If you're operating in a regulated industry and wrestling with how to adopt AI without surrendering your data, your context, or your accountability — I'd like to compare notes.
Own the data. Own the context. Earn the trust.
Every enterprise is being sold the same story: hand your data to a model provider, wire up an API, and transformation follows. I think that story is backwards.
Models are becoming a commodity. Context is not. The enterprises that win with AI will be the ones that own their data and their context — the domain knowledge, the edge cases, the regulatory constraints, the institutional judgment — rather than outsourcing them to the model providers. Context is the asset that compounds.
Owning context is an architecture, not a slogan. Generic AI fails in regulated work because regulated work is defined by its exceptions. What's needed is an industry harness per vertical: domain-specific models, evaluation benchmarks that reflect the actual work, governance built into the system rather than bolted on, and humans in the loop — not as a compliance fig leaf, but as the mechanism by which the system improves.
And trust — not labor — is the adoption unlock. Most consultancies speak the platform language but still run labor-led AI projects, because that is where this quarter's revenue is. The differentiation is trust plus reusable, industry-specific capability: set the right AI foundation, build the harness for the vertical, then launch, rinse, and repeat at enterprise scale.
Nobody says operations roles will go away… it will be more complex and knowledge-based. It's the right evolution of a human skill set with technology modernization. That's how we need to upskill ourselves and our workforce. — On Conduent's main client stage, 2018 · six years before "augment, don't replace" became consensus
Seventeen years in consulting, eight operating a $6B public company's technology, a CEO seat — and now a founder's desk. Every number below is evidence, not identity.
Self-improving, human-in-the-loop AI for regulated industries — the thesis, built with my own hands.
Ran the $400M technology-services P&L; returned it to growth after ten quarters of decline.
Built the technology organization of a ~$6B public company from zero, then rebuilt its product portfolio into a platform business.
Turned around an $800M enterprise-applications P&L while integrating IGATE into a company fifteen times its size.
Ran a $100M financial-services applications P&L and grew it 15% in a year.
Twelve years, programmer to senior director — built a CRM practice into a $100M P&L and ran enterprise applications for Capital One, Fidelity, Goldman Sachs, Citi, and American Express.
Delivered Mahindra's first enterprise CRM implementation — and the lesson that still holds: transformation is 30% technology, 70% people.
Six of the flagship builds. Context, the move, the number.
Launched five enterprise platforms on Microsoft Azure (2019): Dara conversational AI, Analytics, Automation, Blockchain, Mobility — strategy unveiled on stage, announced to the market, shipped as product. Quoted by name in the public announcement.
Consolidated 300+ bespoke platforms into 100 standard products, unified six acquired business units, raised product reuse from under 10% to 70% — ~$400M saved while technology-led revenue grew from $800M to $2B.
Inherited a crisis: 102 data centers, 120 Severity-1 incidents a month, a three-year PCI backlog. Delivered 20 data centers, under 5 Sev-1s a month, PCI in six months, and cloud adoption from under 10% to 70%+.
Real-time, ledger-based payments moving more than $1B a year across 22 states — SNAP, EBT, and unemployment disbursements. Irrevocable money at government scale.
Built and commercialized a net-new Data & AI business inside a $400M services P&L: $15M revenue in 9 months — GenAI for a CPG major, a bio-pharma genome AI hub, a bank KYC Risk Platform.
Product-led sales won a $1B TCV transit contract — the largest in company history — built on IoT tolling that cut roadside installation footprint, time, and cost by 50%.
Writing on enterprise AI, modernization, and the economics of technology — plus keynotes delivered to Fortune-500 clients and industry analysts.
The difference that matters, a four-step adoption playbook, and use cases across healthcare, financial services, and beyond.
Five ways AI extends the life and value of the systems you already run.
From batch to instant and irrevocable — written from experience moving $1B+ a year.
A thought leader with great business and technology acumen and a leader who can drive great outcomes. I have learned a lot from him, enjoyed doing business with him. He will be a tremendous asset to any team.
An exceptional and dynamic leader who combines deep technical knowledge with a strategic mindset. His ability to quickly course-correct an extremely large, troubled product and technology delivery organization makes him a true asset to any organization needing to transform.
Empowering his team, letting them take the bold steps, and then standing behind them like a pillar, are his core ethos. He has orchestrated some unimaginable deals — and continues to build solutions focused on the pain points commonly seen in the industry.
His leadership and strategic vision have been instrumental. Beyond his professional acumen, Rahul is a compassionate leader who genuinely cares about his team's well-being and development. His dedication, expertise, and people-centric approach make him an asset to any organization.
Alongside building Kaia AI, I take on a small number of engagements where the operating history is directly useful.
Technology, product, and AI strategy for boards and CEOs — especially platform transitions, AI adoption in regulated industries, and technology-organization turnarounds.
Operating-partner and diligence work: product and technology assessment of software and services assets, value-creation plans, first-hundred-days execution.
Talks for client, analyst, and investor audiences on operationalizing AI in regulated industries, the services-industry AI transition, and platform economics.
If you're wrestling with AI in a regulated industry, evaluating a technology organization, or building something that needs to be trusted — I'd like to hear about it.