Email, politics, machine learning, and the color purple
The idea hit me like a brick. I stopped hearing the cheering. I stopped even being in the crowd for a moment. Someone was asking me a question and I wasn’t paying attention. All I heard was white noise.

I still can’t believe it’s been three years since this odyssey started. Time has the fascinating capacity to dilate when the world’s moving fast around you. This dilation especially hits hard if you work from home full time. Everything becomes relative in your own little microcosm.
I’ve been quietly building software at the edge of obscurity since George W Bush took office as the President of the United States. I was building apps before they were called apps, when binary sizes had to be below 64k, BREW conferences were a ‘thing’, and network speeds still couldn’t stream video. When most contemporary founders were finishing their CS degree, I was hustling ringtones and touching up wallpapers for Source magazine. My entire career has revolved around the growth of the mobile market.
If you’ve used a Yahoo product in the last 7 years, I’ve likely worked on it, or worked with someone who worked on it, in some capacity. That weird sound Yahoo Mail on Android makes when you have a new email? I made it. You have me to blame for it.
I likely won’t build the next snapchat, and I don’t make reusable spaceships, but I’ve got one hell of a story to tell about my time at Yahoo and how it relates to where I am today.
In 2010, I was desperate for a job. I’d been stringing along bootstrapping a gaming startup with some friends in Minsk by night and working for my old quality assurance manager, Trevor Ridgley, during the day. I’d made about 35k that year and somehow still lived in San Francisco. I was running out of money. I applied to every product manager job I could find, hoping for at least one callback. I didn’t care what I worked on. I knew if I could get in the door, I could sell them on me. My girlfriend (now wife) was letting me sleep on her couch and supplying me with a steady flow of Popov vodka and cheap cigarettes. I was living off corner store chicken wings and packaged EZ burgers in the Tenderloin. I was near the brink of homelessness. I was totally freaked out. I wasn’t sure if I was going to recover from the 2008 crash. Shit was bad.
My ‘pay as you go’ Virgin America T9 burner phone rang. It was Yahoo of all companies. They wanted me to come in and talk with Erik Collier and Sean Pudney. I bribed Sean Pudney with a bottle of Scotch he liked at my interview after internet stalking him. Erik Collier let me talk long enough to convince him I was totally qualified to work on a web property (which I wasn’t).

Erik had me working on Yahoo Groups. Just the settings pages at the start. Groups was my introduction to working on email products. Funny aside; Yahoo Groups is one of the most engaged with email types on Yahoo’s network. Yahoo Groups emails have a material effect on Yahoo’s bottom-line. If you’re a Yahoo Groups user, you’re one of their most valuable users. Yahoo Groups users also like to send Yahoos who work on Yahoo Groups some pretty detailed death threats. Thanks again, Yahoo Groups users out there, for doxing me and faxing the board, demanding my entire team be fired for a botched interface redesign.
I digress.
Working on Groups was where I started to deeply think about email as a form of communication: how people used email, what kinds of email people opened, what percentage of email was from people vs robots, how to detect spam and automated email. The data behind email was intoxicating; it’s hard to describe. We’re talking millions of data points. Looking at email data is like looking into a clear night sky.

Fast forward a year, a few properties later, and a few mass layoffs; I’d ended up working for Chris Yeh, now SVP at Box. Chris was one of the best mentors I’ve had professionally. He had this great philosophy about guidance: No babysitting. He’d let me run as far and long as I wanted. Take on whatever I wanted. His form of management felt more Hunger Games than HBO Silicon Valley. I’d call him punk rock if he didn’t play in a local symphony. Chris taught me trust is a valuable commodity and faith in those you manage matters.
Chris and I were in charge of Yahoo Contacts and a few other properties at the time. Little known fact: Yahoo Contacts is one of the largest repositories for personal information on the internet. We are talking billions of address books. Both Linkedin and Facebooks graphs would not exist without it. It’s that big.
Contacts was my crash course in machine learning. We had some extra funding for a small project to identify groups of people you emailed with most and suggest names for these groups based on what you talked with them about. Something similar to Google’s defunct Circles product. Coincidentally, it was called Yahoo Circles. It was pretty awkward when Google Circles released and I was half way through getting branding approval. At the time I’d made a name for myself internally as ‘the groups guy’. I had a four man team on it and predominantly all the work being done was by a guy named Dayong Huang. Dayong is brilliant. He’s quiet, but when he says something he’s usually trying to make a point.
Dayong, and a few other engineers, helped educate me about machine learning and how it could be applied to complex problems. He helped form the basis of my understanding of natural language processing and machine learning. Want to get into solving big meaty human problems? Start with machine learning.
Around this period I was introduced to David McDowell and Lee J Parry. Contacts got me interacting with the Mail team proper. I say ‘proper’ because of the odd corporate landscape at the time, Chris and I were divorced from the David’s Yahoo Mail desktop team. We didn’t report to them. Lee Parry, who also didn’t report to David, was off managing Yahoo’s big mobile apps riding the first wave of smart phone adoption. Everyone’s goals didn’t always match up. We always had conflicting agendas based on competing company objectives. We made it work somehow.
This is when my career in email took a quantum leap forward.
Lee Parry. Lee was short, thin, well spoken, well dressed, British and had a great accent. He was virtually untouchable politically after his mobile team successfully translated both Yahoo Mail and Messenger onto the newly minted iPhone App Store. He was popular and well liked internally. He’d go on stage every few weeks to talk about our mobile future and you could see a physical positive change in the room. He wasn’t a suit. He had heart. He felt authentic. Best of all he sounded like he cared. He was the visionary Yahoo needed for a post desktop world.
After another brutal layoff purge, there wasn’t any other PM left that knew anything about platform work on the mail team. I’d transitioned into working with an amazing group of server engineers, lead by Krishna Penmetcha, trying to modernize their development practices and and keep the email servers alive. It was a unique opportunity for me. I enjoyed solving problems that only services at that scale have. I got close with my engineers. We were a family. (Am I right, Vivek? Miss you Nitu.)
Running a platform also came with a specific set of corporate political benefits. I was now in every meeting. I knew the email road map. I saw where we were going and had a voice in what that future was at Yahoo.
Things got really political from here on out. I’d developed into the equivalent of product manager ‘delegate’. Chris would send me off to work with foreign product groups internally as a representative of the mail platform. Homepage wants something? Send Russ. Autos? Games? Send Russ. At the same time, the divide between desktop and mobile email started to widen. As mobile computing came to swallow the world whole I found myself interacting with Lee more than David’s desktop team. Lee’s charisma has a habit of being infectious. That, and we both were the only PMs on campus who smoked. Most of our meetings took place at the edge of campus while chain smoking.
This is where I hit a snag in the proverbial river, getting caught up in a quest to do something that, from the start, didn’t even sound feasible. A dark turning point in Yahoo’s eventual trajectory towards ‘Oath’. We joked internally the project would out live us. It almost did. It would take one of Marissa Mayer’s first acts to save us all and one of the many steps she took to save the company.
It started with a conversation. I’d been pitching an ephemeral, small group chat app based on the Mail platform to the company for some time. We even had a prototype working. Lee had caught wind of the prototype. Both of us knew the fabric of future communication was messaging. He’d also been pitching a group chat application on top of the Messenger platform, but didn’t have a prototype. Lee confronted me one day as I was walking back to my desk from lunch. I remember I had eaten chicken stir-fry that day. We had a heated argument about what both us were proposing. I was unprepared; Lee had caught me off guard. It became quickly apparent that my idea had to die for Lee’s to exist. There was no middle ground. The same basic idea, but two very different organizations technically vying for who would build it and what technologies would power it.
Yahoo knew the future of media consumption was in messaging. We were talking about it internally in 2010.
Several days after my verbal conflict with Lee, I got a calendar invite about a new group messaging service Yahoo was building (later called Yahoo Hub). In those meetings, I was given three options:
A) Work for Lee on his new group messaging initiative bridging the gap between email and messenger built on top of the Yahoo Messenger platform.
B) Work for David on desktop mail and give up on messaging.
C) Get fired.
After weighing my options, I threw my lot in with Lee. I was interested in group messaging at the time and working for him was better than looking for a new job in an unstable market. I was still recovering from multiple mass firings like I was one of the final contestants on Survivor. My first responsibility was answering the following question proposed by Lee: “How do we make Yahoo Email and Yahoo Messenger seamlessly talk to each other?”
We spent a year and millions of dollars trying to answer Lee’s question.
Lee and I became inseparable over time. We clicked. The team grew as did the size of mobile user bases. Our small corner of the Sunnyvale office was becoming the focus of the company’s future.
It was 2012. Marissa Mayer became Yahoo’s CEO.

When Marissa hit the floor, the entire tone of the company changed. She brought hope after years of anemic leadership. We believed in her. Her first few weeks she went through almost every single major project currently in development by hand, face to face, and evaluated if said project should exist anymore. Sitting in a room with her as a newly minted CEO was one of the more terrifying moments in my professional career. She’s smart. Scary smart. Being around her was humbling.
Week two into Marissa’s rise to power, she gingerly brought me and a few other technical leads into a room. She wanted a demo of Yahoo Hub. We spent about ten minutes huddled around a table pitching her on the idea of messaging connecting all of Yahoo’s product offerings and showing her our first stable iOS build. We went on to explain all the hard work we’d put into it and how we saw it effecting the business going forward. We were working under the assumption Hub would be the connective tissue of a whole new mobile first Yahoo. After the demo was done, and a long awkward pause, she calmly instructed us to stop development of the entire project. Full stop. The room went dead silent. You could hear the AC running. We’d all spent a year of our collective lives on trying to build this hybrid email/messaging platform. Millions of dollars. Hundreds of man hours. We were betting the company’s future on this platform. She’d just walked in the door and smashed it, like a rotten pumpkin being cracked open with a hatchet. That ten minute demo felt like a lifetime.

I am not sure, and someone can correct me on this, but I may hold the dubious honor of having the most projects Marissa Mayer cancelled without getting fired myself. I’m not proud of this statistic, but I am happy she liked keeping me around.
In hindsight, I believe Marissa made the right decision cancelling Yahoo Hub. She always had a knack for making the hard decisions. It takes a tough person to do what she did. Most people in the valley would give up and go live on a desert island instead of make some of the choices she had to make during her tenure as CEO of big purple. I could go more into technically WHY I think we made the right decision, but that could be a blog post in itself. We were building on quicksand and gambling with the companies future with one app. All the eggs in one basket. If we’d pushed forward there may not have been a lot left of Yahoo (Oath) at this point. She may have ‘saved the company’ with that single decision.
Marissa gave the team a new charter. Update all Yahoo Mail clients on all platforms at once with modern interfaces. Up until then we’d been releasing new versions of mail haphazardly across web and mobile. Sometimes new features would release on one platform and it might take months for the others to gain similar support. We thought it was impossible at the time to sync releases. None of the major product groups had ever done a cross platform timed release. Everyone worked day and night on the revamp. Every floor. Every mail product team all awkwardly shambling in the same direction. The general consensus was if we couldn’t make it happen, none of us would be around in six months. The threat was palpable. It felt like Marissa’s first test to see if she was going to shit-can all of us for incompetence.
Nothing is better than the feeling of being an underdog. That’s what being a Yahoo employee feels like. Like you’re on a baseball team that hasn’t won a game in 10 years. In their purple hearts, every Yahoo employee knew they were the underdog. Every day you show up and try knocking a metaphoric ball out of the park when everyone expects you to strikeout.
Three months after Marissa’s mandate, we simultaneously deployed a unified Yahoo Mail interface across Android, iOS, Windows 8, and mobile and desktop web. It was surreal. Execution like this was unheard of internally. Yahoo had never done anything like what we were doing. Operationally we were Yahoo at it’s best. Tech media didn’t even notice. We were just another purple blip on their radar.
Lee found me a new home managing Yahoo Mail on Android, giving me a real challenge after Yahoo Hub got shut down. It was simpler. More contained. A large user base and a decent interface. I had a great engineering team lead by Stephane Karoubi, now a director of engineering at Apple.

Life got a lot less volatile after that. Marissa had an entire floor built for us with a modern layout. We were always hiring. New faces started to show up everyday. It felt like the company was really turning around. There was an energy in the air. I’d call this the golden era of my stay at Yahoo if it wasn’t marked by tragedy. I still have a hard time even thinking about it now, but it’ll make sense within the context of what came next in my career. Writing about it feels therapeutic.

Lee got sick. He had a traumatic brain injury, called an AVM. It’s congenital and only surfaces in your 30’s. There was nothing he could have done to stop it from occurring. No lifestyle changes could have stopped it. God, or whatever is out there, can be cruel.
I took what happened to Lee professionally and personally hard. I was one of the last people he talked to before surgery to repair the AVM. I love Lee like a brother. I’d always followed his lead. We knew how to dance. I’m pretty sure his desk is still untouched as memorial to all he did for Yahoo. I don’t think people realize how much he changed that company. Lee was Yahoo. I went to a pretty dark place mentally. I acted like he was on vacation for a while. It was all that kept crippling depression at bay. I still miss building software with him every day. Like I said, we just worked.
With Lee retired, a power vacuum was created. We had all the users, all the engagement, and drove all the mobile traffic. We were a mack truck barreling down the information super highway. The future of the company was inside our apps. We held all the keys. The leadership team was stunned by Lee’s departure. There wasn’t a game plan for Yahoo mobile without Lee. They didn’t hire anyone to replace him. We had a product team without a leader for almost 3 months. Without really asking permission I stepped into Lee’s very large shoes out of necessity. We needed leadership; people needed yearly reviews, QPRs, and bonus negotiations. He didn’t work this hard for it all to go to shit. Some of my coworkers viewed it as a coup.
I had no idea what I was doing at first. I had no idea how to manage people. I had no idea how to present to thousands of people. Public speaking wasn’t in my skill set. I was suddenly in charge of the most important product the company offered, shepherding some of their most valued customers. I know it’s cheesy, but Dune got me through these months. I just kept telling myself every morning, “Fear is the mind-killer. Fear is the little-death that brings total obliteration…I will permit it to pass over me and through me. And when it has gone past I will turn the inner eye to see its path. Where the fear has gone there will be nothing. Only I will remain.” Thanks, Frank Herbert.
There’s a lot of stories between taking on Lee’s role and where I’m going to pick up. If I tried explaining all of them this post would turn into a novel. Ash, Chloe, Drew, Julianne, John, Andreas, Andrew, Albert, Rossi , Sean, Harshad, Eileen, Satyadeep, Gopal, Alicia , everyone I missed — If I ever decide to write more about these experiences, I’ll get to you. We had a wild ride.

A few years went by and I became a local fixture of the company. I presented to Marissa regularly. I went on stage and talked about our shared future. I participated in shaping the new #400090 hex code Yahoo. It was powerful stuff. My beard started turning gray.
Relationships with the desktop mail team improved thanks in part to the injection of the Xobni team who was now running it. Josh Jacobson and Jeff Bonforte became staples around the office and even started to help inform where the mobile mail client was going. I brought on people from the desktop team into the mobile unit to help heal the rift between. Suddenly we were all working towards common goals. It was a pretty exciting time to be alive and working at Yahoo. We had legitimate aligned ‘synergies’ and everyone gave a shit.
Jeff Bonforte made a huge impact on me. He is charming, goofy, warm with technical chops. He planted the seeds of what later would become the focus of my professional career. I can’t properly articulate how much he shaped where I am today. Never underestimate the impact a five minute conversation can have on someone looking for guidance. Five minutes can change someone’s life.
I crashed one of Jeff’s staff meetings as curiosity got the better of me. I wanted to know more about him and we crossed paths rarely other than on stage. Characteristically, he was all over the place and excited about data. Data, data, data. What stuck with me was a moment where he excitedly interrupted his perplexed staff, after looking out an A4 window for 10 minutes, “Email is the biggest gold mine ever! There’s so much data and so much to be learned!” I agreed with him silently observing from the corner of the room. Wheels spinning in my head.
It was late on a Friday in July. We’d just launched Yahoo News in Yahoo Mail. I remember stepping up onto the stage after a couple of beers. I never could get comfortable with getting up in front of the company. I always got the shakes. Most times I just blacked out and went into performance mode. I’d been asked to announce the metrics for the new mail news and weather features. I’d been coached all morning by the company’s professional trainers. The lights were bright and I was already sweating. My heart was pumping pure adrenaline. Who gets this excited about talking about growth numbers?
This is where I found my gap in email.
I went on to pontificate via slick slides (Thanks Guilherme Neumann) we’d increase engagement by 70% per user. Average time spent was up 65% per user with a 17% QoQ increase of active users. Everyone was cheering and clapping. I still have no idea what exactly I said, but I am pretty sure I hit all the marks. I got a pat on the back as I was walking off stage. More engagement meant more revenue. More engagement was good for Yahoo. More engagement was good for share holders. Something sparked in my brain. My mind started to race. It was like I’d been staring at a painting for a little too long. I’d found a hole in email and I could see its edges crisply mentally.
You have to really ask yourself what ‘user first’ means. When I was walking off that stage I started to ask myself; “Why are we cheering?” Does more email engagement mean we are doing the right thing for users? Who really benefits from more time spent in an email inbox? The user? Yahoo? The advertiser? What do we give users for this additional time? What benefit do they get? The answer became clear to me and for a moment nothing around me really mattered anymore. Nothing. Launching a Yahoo News tab in email did nothing for user. It wasn’t ‘user first’. It was so we could make more revenue per user. Money, money, MONEY. As the seed of my future endeavors was planted, I started to ask myself a secondary series of questions. A rush of thoughts came to me. The idea hit me like a brick. I stopped hearing the cheering. I stopped even being in the crowd for a moment. Someone was asking me a question and I wasn’t paying attention. All I heard was white noise.
What would an email client look like if it was designed to make a positive impact on the user’s life?

By mid 2014, something curious happened. I’d seen a massive empire formed. My small team of 7–8 mobile engineers had grown into a floor of over 300 engineers under Marissa. I was managing two products and had a team of 4 product managers I was grooming. Our numbers were spot on. The world had changed thanks to Marissa. What I didn’t realize was the company had also changed into one where I’d been rendered obsolete. Something the New Yahoo called “Old Yahoo!” a constant reminder of an embarrassing past. We were jaded and cynical. Damaged goods. We didn’t exhibit newly required ‘relentless positivity’ expected by the New Yahoo. Filo, I now know how you felt sometimes. I get now where you were coming from.
My new counterparts were all Google and Amazon alumni. Hired at a premium over the last few years. Fed on new policies of free food, free phones, and dotcom luxury. A new ‘breed’ of product thinkers. New blood. Ex-CEOs with million dollar exits and 18 year old wunderkind. No one was comfortable the largest mobile product was being helmed by some 29 year old with no college degree and nothing on my resume that screamed Innovative Visionary. I didn’t fit into the new narrative. My manager informed me he’d decided to break up my role into new positions managed by Google luminaries and acquired talent. I’d simply survived long enough to become an awkward reminder of an old company desperate for talent willing to hire anyone that remotely looked good on paper. It’s a weird feeling when you realize you’re no longer needed by a corporation. It was heartbreaking.

These were the times you do some soul searching. I had poured my entire life into Yahoo. I ate, slept, and bled purple for half a decade. I started by politely asking a few people I looked up to internally if they’d get coffee with me. Jeff Bonforte sticks out as memorable. I appreciate how candid he was at the time. He had one simple recommendation: “Go. Build. Something.” Not at big purple. Something outside of big purple. Easy to say, but the way he said it had impact. I took his advice to heart.
One of the great things about working at a big company is you get to see where the gaps are in their portfolio. Where were those gaps and what could I build there? Where could I explore something new? Investing time at a large company can open your mind to a world of problems you have the unique ability to solve the day you walk out the door.
My mind went back to stepping off that stage in July. What if I could build an email client designed to improve someone’s life? What would I build? Where would I get the money, the people, and time? Would it even work?
I updated my Linkedin and optimized with SEO tools. I paid a premium to make myself appear more frequently in search. I wasn’t sure why I did it, but I knew if I waited long enough something or someone would notice me. I had faith the world would bring the answer to me. Call it a moment of cosmic intuition. I knew something was out there looking for me.
I got an email a month later from Dane Baker. He was starting an applied sciences startup trying to explore how machine learning could be leveraged to make email smarter. He was looking for someone with consumer email experience. It was entirely privately funded. For a moment I didn’t know if he was a scam or the real deal. Taking a risk I replied.

Over a month, Dane, the team, and I bounced ideas off each other over personal email. We came up with a name for the company. We’d call it ‘Codeq’. Keys, Codex, it all felt like a good play on words. I still wasn’t sure if he was serious until he offered me a term sheet. I’m not an innately spiritual person, but to this day I think there had to be some kind of force to bring Dane and I together at that exact moment with the exact same goals. We had a dream.
We were going to build an email client leveraging machine learning to save users’ time.

I took a month off to decompress. I hadn’t taken a vacation in so long I’d accrued 30+ days of PTO. Got married to my lovely wife and travelled the world for a bit. The day I got back, I turned in my laptop, left Yahoo, and never looked back. Off to build my own future. A big bold idea just like Jeff recommended. A user centric future with my new partner in crime, Dane. We were setting out on an adventure to rethink email for the better. An email service for humans and not advertisers.
As a team, we’ve spent about 2 years perfecting our new email app, called Courier, using some of the most advanced machine learning techniques ever developed for email. Built for the little guy! The person out there wishing they’d spend a fraction less time reading email and more time paying attention to their son or daughter.
I can’t wait to tell you more about it.

This is one of a series of posts about my life, Codeq, and Courier — our new email app! Please join our beta list and help us save the world one email at a time. On average, we think our email app can possibly save you about 15 hours a week by summarizing your email for you. Best of all? It is free.
