Presenters

Source

From Boat Person to Tech Titan: Tuan Pham’s Epic Journey Through Uber’s Hypergrowth and Beyond 🚀💡

Ever wondered what it’s like to build a tech giant from the ground up, navigating chaos, impossible deadlines, and unprecedented scale? Our latest conversation with Tuan Pham, a legendary figure in the tech world and former CTO of Uber, pulls back the curtain on this incredible journey. From fleeing Vietnam as a child to leading one of the most complex engineering organizations ever built, Tuan’s story is a masterclass in resilience, innovation, and leadership.

A Journey Forged in Fire: Early Life & Lessons Learned 🔥🛤️

Tuan’s story begins not in a classroom, but on a “rinky boat” escaping post-war Vietnam as a child. His mother’s bold decision to seek opportunity for her sons led them through perilous seas, surviving against less than 50% odds. Arriving in the US with no English and no money, Tuan built his life from scratch.

His entry into tech was serendipitous, sparked by an IBM PC and a friend in high school. He quickly discovered a natural algorithmic thinking and a dislike for doing the same thing twice – making programming a perfect fit. He taught himself BASIC, volunteered at a government agency, and automated financial accounting, reducing a 3-week quarterly workload to just 3 hours. This early success propelled him to MIT, where he truly delved into computer science fundamentals.

After MIT, Tuan worked at pioneering companies like HP Labs, Silicon Graphics (SGI), NetGravity, and VMware. These experiences taught him invaluable lessons:

  • The “Unsatisfied” Researcher: At HP Labs, he worked on cutting-edge medical informatics (networked physician workstations, drug interaction knowledge bases) but found it “unsatisfactory” when research wasn’t productized. He craved seeing his code in users’ hands.
  • Ahead of Its Time: At SGI, he prototyped interactive TV with video on demand and online gaming for 4,000 homes in an 18-month trial in 1994. While visionary (even Michael Jackson and Steven Spielberg saw demos!), it fizzled out due to exorbitant costs ($45,000 per set-top box), teaching him that technology alone isn’t enough; timing and economic feasibility are crucial.
  • Growth vs. Profitability: At NetGravity, an early internet advertising startup, Tuan and a colleague launched the first dynamically targeted ad on Yahoo!. However, the company focused on profitability while a competitor prioritized growth, eventually being acquired by Google. This taught him that sometimes growth, even without immediate profit, can win the market.
  • Dot-Com Bust Resilience: He navigated the dot-com bust, a “scary time” with layoffs and bankruptcies. His takeaway: talents are always marketable, especially if one continuously invests in skills and avoids complacency.
  • VMware’s Game-Changer: Joining VMware when it was a 40-person organization, Tuan helped build its first product suite, VirtualCenter, which powered the cloud platform. He witnessed virtualization become an industry-changing technology, particularly with “vMotion” enabling seamless migration of virtual machines.

Tuan’s career also revealed a pattern: when things got too comfortable at large companies, he sought smaller, more challenging environments. This drive for constant growth and impact would define his next, most iconic role.

The Legendary Interview & Uber’s Early Chaos 🤯📞

After 8 years at VMware, Tuan felt it was time for a new challenge. His strong reputation, built by “doing a really good job at every company,” led Bill Gurley of Benchmark Capital (an early Uber investor) to him. Gurley remembered Tuan from NetGravity a decade prior!

The interview process with Travis Kalanick, Uber’s co-founder, was legendary: 30 hours spread over 2 weeks. Travis, deeply engineering-minded, wanted to “vest” Tuan’s fit, effectively simulating what it was like to work together under pressure. He had a whiteboard list of topics, from hiring and firing to code quality and design, and even a list of five desired engineering team cultural traits. Travis’s passion was undeniable; he once rescheduled a flight mid-conversation to continue the interview! Tuan realized Travis viewed Uber’s business as powered by two engines: operations and technology.

When Tuan joined as Uber’s first CTO in April 2013, the company had 40 engineers, processed 30,000 rides per day, and its system crashed multiple times per week. He had 5 months before the dispatch system would hit a “brick wall.”

Scaling Through Chaos: Rewrites, Microservices, and Internal Tools 🛠️🚀

Uber’s growth was “unusually violent,” forcing the engineering team to “race against time” to prevent system collapse. Tuan’s core philosophy was to “see around corners” and “live to fight another day.”

  1. The Dispatch Rewrite: The initial dispatch system, built with Node.js, was a single-threaded monolith that relied on vertical scaling (moving to larger machines). Tuan quickly identified a “brick wall” coming in October for New York City. He challenged his 3-4 person team with two simple, yet powerful, requirements: a city must be powered by multiple boxes, and a box must power multiple cities (N by M scaling). With “no new features necessary,” the team rewrote the system in a “really scalable way” and deployed it by August/September, just in time.
  2. The Microservices Explosion (Project Darwin): Uber’s famous 5,000 microservices didn’t happen by design, but out of necessity. The backend API monolith was a bottleneck for speed. Tuan declared: anything new must be built outside the monolith. A dedicated team (Project Darwin) worked to decompose it. While in isolation this could take 3-6 months, it took two years because business growth (new cities, products like UberX) meant teams kept adding features to the monolith faster than it could be decomposed. Tuan explains, “no one should be blocking anybody else.” This “chasing our own tail” led to a bulging monolith, but ultimately allowed Uber to “fan out and solve every problem all at once.” Today, Uber has refined this, reducing its microservices to around 4,500 in 2026, even with increased complexity.
  3. Building Internal Tools: Uber initially relied on open-source solutions like Redis and Postgres. However, their “unusually violent” scale pushed these tools “to the breaking point.” A “very painful example” was Postgres randomly failing at scale, with no one in the open-source community able to diagnose the kernel-level issue. This led to building their own solutions:
    • Schemaless: A custom data store.
    • T-Channel: An RPC protocol.
    • Ringpop: A consistent hash ring library.
    • Jaeger: A distributed tracing tool (later open-sourced).
    • M3: An observability platform (later open-sourced).
    • They even moved to MySQL as a table data store, building all logic on top to control their own state and features. This was a direct response to the immaturity of open-source solutions at Uber’s scale in 2013-2016.

The Impossible China Launch & Organizational Innovation 🇨🇳🤝

In late 2014, Travis Kalanick issued a “craziest thing we’ve ever done” challenge: launch in China in two months, requiring services to run on China soil with completely partitioned data and controls. Tuan’s TPM team scoped it to six months (industry friends laughed, suggesting 18 months), but they negotiated it down to four months. Even then, they slipped to five months.

Travis agreed to an incremental launch, but insisted on starting with Chengdu, the biggest city. This “brilliant” strategy meant “doing the hardest thing first” built immense confidence, making subsequent city launches routine. The project pushed everyone to “redline” themselves, fostering a “fearlessness and risk-taking culture.”

Beyond technical scaling, Tuan championed organizational innovations:

  1. Program & Platform Split: With 40 engineers and 3 product managers in 2013, growing to 100 engineers and a dozen product managers by year-end, functional teams became a bottleneck. Tuan, Travis, and Jeff Holden restructured into cross-functional teams (programs for end-user features, platforms for shared tools). They identified 17 business areas, initially funding seven with partial funding for four more. This ensured teams had all necessary skill sets to “just get it done.”
  2. L5A/L5B Levels: Tuan introduced L5A and L5B levels for senior engineers to provide a sense of progress during the five-year journey to staff engineer, acknowledging that “not everybody can make it to staff.”
  3. Easy Internal Transfers: To retain talent and foster growth, Tuan created an “easy internal transfer process,” eliminating the need for manager permission. He argued, “Why do we make it harder for ourselves when our own engineer goes from team A to team B have to ask for all these permissions where they don’t have to ask if they interview outside?” This radical move also incentivized managers to develop and retain their talent.

Leadership, Legacy, and AI’s New Frontier 🌟🤖

Tuan describes his time at Uber in “three tours of duty”:

  1. Fixing the broken stuff: Making systems reliable and rebuilding foundations.
  2. Scaling worldwide: Massive growth, including the China launch.
  3. Navigating turbulence: Staying through Uber’s challenging 2017 period, including Travis’s departure and the IPO, feeling he “owed it to oursel and we owed it to everyone.”

He eventually left Uber in 2020, not due to COVID-19, but a personal realization: money no longer mattered, and he sought roles where he loved the mission, made a big impact, and enjoyed his colleagues’ company.

His “retirement” was short-lived, as COVID-19 grounded travel plans. He became involved with Coupang (an e-commerce giant with 5-hour delivery in Korea) and joined the board of Nubank (Latin America’s largest digital bank). These experiences reaffirmed his belief in “solving the right problem at the right time” with “phenomenal cultures.”

Currently, as CTO at Fair, a B2B marketplace empowering local businesses, Tuan is tackling the most exciting challenge: AI.

  • Boosting Effectiveness: Fair uses AI to boost everyone’s effectiveness and productivity.
  • Product & Search: AI enhances search and recommendations, helping businesses discover products that sell well.
  • Engineering Productivity: They are using swarm coding (orchestrating multiple AI agents) to dramatically increase engineering output. Early adopters have doubled their output, demonstrating that “the output is dramatic.” The next frontier is leveraging AI to build new features on older, complex codebases.

Tuan believes AI is rapidly changing software engineering, abstracting away low-level details and elevating the playing field. However, what makes a great engineer remains constant: curiosity, fearlessness, willingness to innovate, and a drive to break new boundaries. “Complacency is death,” he warns. AI is a powerful tool, and “you can wield the tool properly [to] do extraordinary thing versus you just merely use a tool in a mundane way.”

The CTO’s Blueprint: Build Teams, See Around Corners 🎯🔮

For Tuan, the most important job of a CTO involves two key areas:

  1. Build a High-Performance Team: Focus on talent density, fostering trust, alignment, and a culture where “team A will just want to hire more A-level players.”
  2. See Around Corners: Envision the future 18-24 months out, anticipating business needs, architectural challenges, and talent requirements. “Skate to where the puck would be,” he advises, empowering his team to focus on the 6-month horizon while he looks further ahead.

His advice for young engineers:

  • First 5-10 years: “Find opportunity where you learn the most that push you the most.” Develop skills rapidly.
  • Next phase (Senior/Staff): “Seek opportunity where you can make a big impact.” Leverage your knowledge.
  • Later phase (Principal/Leadership): “Learn to give back,” coaching and developing others, leading with responsibility.

Tuan’s career is a testament to building genuine relationships. Engineers from VMware came to Uber because they “still enjoy working with you.” His philosophy on “behaving work in the perspective of death” emphasizes humility, being constructive, and leaving a lasting positive impression. It’s not about networking hacks, but about altruistically doing good work and being good to people.

Tuan Pham’s journey reminds us that in the relentless pace of tech, resilience, a hunger for learning, and a commitment to people are the ultimate drivers of success.

Appendix