The hinge in Jack Singh’s career was not a pitch deck. It was an Excel sheet opened after a long conversation in a parked car, rain pounding outside. Near the end of five years at Amazon Web Services, Singh told a mentor about a goal he was chasing. The answer was blunt: working for somebody else was unlikely to get him there. He went home, modeled both futures and studied the arithmetic. The cells pointed toward another startup.
It is an unusually unromantic origin for Avahi, the San Francisco cloud and AI company Singh founded in 2022. There was no napkin revelation, no garage mythology, no claim that fate had sent a push notification. There was a grown operator measuring time against ambition. The calculation worked because Singh had spent years gathering the inputs.
He arrived in the United States at 18 for college and worked two full-time jobs to support himself. After graduating, he joined a startup as employee number five. The company quickly grew to roughly 75 people, and Singh wandered wherever the work demanded: sales, customer support, manufacturing, packaging. Long before he spoke fluent cloud, he learned the physics of a small business. A promise made by sales eventually becomes a box someone has to pack.
The education of a useful generalist
That first company gave Singh range, and range gave him an itch. Around the arrival of the iPhone, he started an app-development business, ran it for several years and sold it to another company in the sector. The experience also exposed a gap in his education. He understood startup improvisation, but not yet the machinery of a Fortune 500 company. So he went looking for the missing chapters.
Microsoft and Dell EMC supplied enterprise experience. AWS supplied the ecosystem view. Singh spent five years developing partner sales motions and watching small consultancies grow. Some began with a handful of people and became organizations with a hundred employees; some eventually sold. The pattern looked repeatable, but one customer gap kept flashing.
There is a quiet advantage in learning both sides of a handoff. Singh had sold promises and watched them travel toward support, production and delivery. At AWS, he could see a similar journey at ecosystem scale: a cloud platform set the possibilities, partners translated those possibilities, and customers judged the result in their own operating environment. Each layer could be locally correct while the whole experience still failed. The founder’s job, in Singh’s telling, was to shorten that distance. Technical people needed enough commercial context to recognize the real constraint. Commercial people needed enough technical fluency to stop selling decorative certainty. Avahi would sit at the seam, where architecture and expectation either shake hands or start blaming each other.
Many partners approached startups with a sales motion borrowed from larger companies. The technical depth could be thin, while the cost structure still carried enterprise weight. Singh’s thesis for Avahi was simple enough to fit on a sticky note: work backward from the customer, lead with technical capability and price the work for the company actually sitting across the table.
“Solve customer problems, earn their trust, and they will continue working with us.”Jack Singh’s summary of the Avahi loop
A score out of ten, please
Avahi’s most revealing management tool may be a tiny moment at the end of a call. The team asks customers to rate the engagement from one to ten. Eight and above is acceptable. Anything lower becomes an invitation to keep talking: what missed, and how can the correction become scalable and repeatable?
Singh has said he “lives and dies” by Radical Candor, Kim Scott’s book about direct, humane feedback. The rating ritual turns the idea into plumbing. Plenty of leaders invite honesty in an all-hands meeting and then make honesty expensive in practice. A number is harder to perfume. The follow-up matters even more: criticism becomes a design input rather than a small courtroom drama.
His view of talent changed in a similarly practical way. Earlier in his career, Singh followed the familiar instruction to work on his weaknesses. Later, as he built teams, he decided the better unit of analysis was the group. People should keep improving, certainly, but a company gains more when one person’s strength covers another’s weak spot. The goal is not a staff of polished Swiss Army knives. It is a drawer where the right tools sit next to one another.
One ecosystem, all the way down
Avahi chose concentration over cloud tourism. The company built around AWS, covering migration, application modernization, data, operations and generative AI. In 2024 it earned AWS competencies in DevOps and generative AI, reached Premier Tier Services Partner status and signed a strategic collaboration agreement with AWS. In August 2025, the companies announced an early renewal of a multi-year agreement, with a shared emphasis on getting generative AI from experiment to operating reality.
This is the unglamorous half of the current AI boom. A demonstration can make a room gasp. Production asks whether the result arrives quickly, respects the data boundary, survives traffic, fits the existing stack and costs less than the problem it replaces. The applause is quieter, but the invoice is real.
A project for live-streaming platform Groopview illustrates the point. Its AI co-host had a timing problem: by the time a complex answer arrived, the live conversation had moved on. Avahi built a dual-model system that routed simple questions to a fast model and harder ones to deeper reasoning. The resulting implementation cut response latency by 80 percent and earned Avahi a place on Inc.’s 2025 Best in Business list for Artificial Intelligence and Data.
The useful detail is not merely that two models beat one. It is that architecture followed the social constraint. In live media, a correct answer delivered after the moment has passed can be functionally wrong. Speed is part of the product, and the business problem gets to set the technical priorities.
When the spreadsheet met a bank run
Founders enjoy models because models keep their assumptions in neat rectangles. Reality has a more improvisational layout. During the 2023 collapse of Silicon Valley Bank, Avahi’s funds were caught inside. A public account from Singh’s communications adviser said he covered employee payroll from his personal money while the company’s cash was inaccessible. Singh also appeared in news coverage of the bank’s failure.
The episode made his favorite abstractions suddenly physical: runway, trust, obligation. It also rhymed with the teenager who had worked two jobs through college and the startup employee who knew that work eventually lands on a real bench. Leadership, on such days, is not a value printed on a wall. It is deciding which account absorbs the shock.
The small-win machine
Singh’s founder advice includes a deceptively modest instruction: “Celebrate small wins.” It suits a business where progress arrives as a long series of migrations, integrations, audits and response-time improvements. There is rarely one cinematic switch-flip. More often, a system gets a little faster, a deployment a little safer, a customer a little more willing to hand over the next problem.
His own career reads the same way when stripped of hindsight. Arrive. Work. Learn the whole small company. Build one. Sell it. Notice what you do not know. Join larger organizations. Learn an ecosystem. Help partners grow. Open the spreadsheet. Begin again.
That sequence explains why Singh talks about adaptability, vision and resilience, but operates through mechanisms: a rating at the end of the call, complementary strengths on a team, a model for the decision, a technical answer tied to a customer constraint. Vision is charming company at dinner. Mechanisms show up for breakfast.
A founder’s grand strategy is often just a collection of small promises that survived contact with Monday.The Avahi story, in one operating sentence
Avahi’s ambition is to make cloud and AI useful to companies that cannot afford an endless laboratory phase. Singh’s route to that ambition has been neither straight nor particularly tidy. It runs through a packaging station, an app shop, several enterprise sales organizations, the AWS partner network and one rain-soaked conversation.
The spreadsheet offered no guarantee. It only made the choice legible. Singh supplied the rest: a company built around the belief that technical credibility can earn commercial trust, one scored conversation and one shipped system at a time.