LEADERSHIP / 01

The judgment
behind the work.

AMBIKA SUKLA / FOUNDER

Prestilla Labs is led by a builder who has operated at every layer of AI delivery—from fine-tuning models and writing production code to founding companies, creating enterprise AI organizations, and governing systems in regulated environments.

The advantage is not a single methodology. It is knowing which problem you actually have, what credible evidence looks like, and what it takes to make the answer survive contact with an organization.

AS
BUILDER / FOUNDER / OPERATORNYC

WHY THIS MATTERS / 02

Strategy is more useful when it is accountable to the build.

Ambika has led AI from the research lab, the engineering floor, the executive table, and the founder’s seat. That range makes it possible to connect model behavior to product decisions, platform architecture, compliance obligations, team ownership, and business outcomes.

OPERATING PRINCIPLEDo not recommend a system you would not be willing to build, measure, and defend.

SELECTED EVIDENCE / 03

Experience measured in systems shipped and organizations made capable.

25+years building technology

From real-time financial systems to frontier AI products.

0→1founder and operator

Research, product, fundraising, customers, and production ownership.

35+AI organization built

Researchers, ML engineers, platform engineers, and technical leaders.

30+AI products governed

Internal and vendor systems cleared for regulated production.

1,000+users for a production financial-services AI agent
50%reduction in ad-hoc data requests through semantic AI models
30%of customer-support requests automated with an AI agent
<30 mindocument completion reduced from a full day

OPERATING HISTORY / 04

Frontier work.
Production stakes.

A career spent making emerging technology useful inside institutions where reliability, security, and accountability are not optional.

012025–2026

CAIS

Managing Director, Head of AI and Data

AI strategy → governed production

Built the AI organization, platform, governance council, and production agents serving advisors, asset managers, sales, marketing, operations, and customer support.

022024–2025

Salesforce

Sr. Director / Lead AI Architect

Research systems → enterprise scale

Set AI strategy and reference architecture across the CIO organization and helped shape flagship agents by improving accuracy, latency, and production performance.

032019–2024

NLMatics

Founder and CEO

Frontier research → durable company

Built early prompt-based extraction and hybrid RAG products for banks, legal firms, and healthcare—with cloud and on-premises deployments meeting SOC 2 Type II and HIPAA requirements.

042006–2019

Morgan Stanley

Executive Director

Executive case → institutional capability

Launched one of Wall Street’s early AI Centers of Excellence, built a 35-person AI organization, and created the firm’s first on-premises GPU-based MLOps platform.

052001–2006

Goldman Sachs

Vice President

Complexity → operational clarity

Built advanced analytics, search, data access, visualization, and real-time market-monitoring systems in a demanding production environment.

THE TEAM MODEL / 05

Senior people, personally accountable.

Prestilla Labs assembles small, high-leverage teams around the actual problem. Every practitioner is personally vetted by Ambika for technical depth, judgment, and the ability to work directly with stakeholders—not hidden behind a delivery layer.

When the right answer is an internal capability, we help organizations design the team, define its charter, hire and assess talent, establish engineering standards, and transfer ownership.

ENGAGEMENT TEAM / QUALITY CONTROLPL–05
01
Personally vetted

Selected for relevant depth, not generic availability.

PASS
02
Builder-led

Senior technical leadership stays close to the work.

PASS
03
Outcome-owned

Evaluation and transfer are designed into delivery.

PASS
04
Capability-building

Your organization becomes stronger after we leave.

PASS
TWO VALID OUTCOMESWe build the system with you—or help you build the team that will.

A NOTE FROM AMBIKA / 06

“The best AI work combines scientific honesty with an operator’s bias for making things real. That is the standard I bring to every engagement—and the standard I expect from every person I put on a team.”
Talk through the hard problem ↗