About me:
My name is Sophia Luo and I am currently an engineering manager and product manager on Scale Rapid, our self-serve and rapid experimentation data annotation product. During my time at Scale, I have been an engineer and product manager on several different teams; including eCommerce, Marketplace, and Platform. In four years, I graduated with a Masters of Engineering in Computer Science, Bachelors of Sciences in Computer Science, and Bachelors of Sciences in Mathematical Economics from the Massachusetts Institute of Technology. In my free time, I like to dance, draw, eat good food, and plan fun and new experiences with friends.
Top Slacker at Scale AI
When I first graduated college, I kept telling my friends how confused I was trying to decide where to go for my first job as a software engineer. Fortunately, I had the opportunity to choose a path that was best suited for me, as I intentionally reviewed multiple offers. This was a major life decision, the biggest one I had to make since deciding which college to attend four years prior.
When making this decision, I accumulated 18 different company traits, including growth opportunity, work life balance, mentorship, and more that I would evaluate to see if the position would be a good fit. In hindsight, the whole process was very heavy handed. In the end, I realized that there were only a few traits that really mattered to me and it became clear that Scale was the best choice:
- Growth opportunity: Scale was quickly growing, and I knew that there would be many organic career opportunities. It also became apparent that the company was open to providing opportunities on important projects to those willing to take them on– regardless of tenure or experience. This was very important to me as someone starting their career. I wanted to be able to dive straight into the deep end and learn by doing.
- Talent: The high concentration of talent at Scale was awe-inspiring. Just from talking to the employees and searching their profiles on LinkedIn, I was both intimidated and excited to join a group of people who were so talented. I wanted to surround myself with and learn from some of the most talented people in the industry.
- Culture: After spending time with Scale employees and also hearing about their experiences, I found that there was a very tight-knit community of friendship and trust. As a fast growing startup, there were ups and inevitable downs, but it was clear that the Scale people had a very tight bond that helped them get through even the toughest of times.
Fast forward about half a year and — boom — the pandemic and quarantine. The last semester of my college career was cut abruptly short, and I found myself back at home feeling sad, bored, and lonely. As a lonely new grad starting at a company where I couldn’t see the faces of my coworkers by default, I regularly sent cold messages to people across the company on Slack and asked to schedule 1:1 meetings to get to know them.
As the number of people I became friendly acquaintances with grew, so did the number of Slack messages I sent. In fact, I was sending Slack messages outside of my normal work duties so much that I found myself added to our #top-slack-users channel after a few months of starting at Scale.
As I was getting to know people through these Slack DMs and 1:1s, I also found myself working with many of them through my first large-scale project. Within my first six months, I was tasked with developing Scale’s first marketplace model as the third engineer on our Marketplace team. Developing this model from 0 to 1 involved applying many econometrics and linear programming concepts I learned in school. I needed to compute annotation pipeline capacity requirements and derive other useful variables. Furthermore, I was learning about data orchestration technologies and systems design techniques to be able to compute across all live annotation pipelines. The implementation and rollout of the model also involved more cross-functional work than I could ever imagine; I was interfacing with product managers, program managers, operators, engagement managers, and engineers across the organization who all wanted to learn about and integrate with my work.
For being so early in my career, I felt blessed with the opportunity: there was just so much to learn and I was excited to be owning such a large surface area of work so early on in my career. I felt like I was learning and growing very quickly. In addition to improving my raw technical ability, I was learning how to become a more effective engineer by learning how to program manage, communicate with stakeholders, prioritize feature requests, and drive outcomes.
As I was working on the marketplace model, I also became heavily involved in our university recruiting program. I was very excited to write, test, and calibrate new engineering interview questions, and it was Scale’s first time courting candidates completely virtually. Instead of hosting in-person events for candidates to learn more about Scale and the culture, we had to be more creative. I ended up organizing and playing countless virtual Among Us games with candidates, and we also threw together a Trivia Night where we invited candidates and Scale employees to compete for fun prizes together.
After quite the eventful first year at Scale, I found myself with yet another career growth opportunity. I had transitioned from the Marketplace team to become the third engineer on the eCommerce engineering team where we were building a new product from 0 to 1 to develop AI-enabled data infrastructure for eCommerce.
Being on the eCommerce team was quite the rollercoaster ride. As a team of three engineers, we first worked on a proof of concept for a customer who is still one of our largest today. This was a two week sprint where we built everything from scratch. With the correct licensing and legal practices, we needed to collect raw catalog data, extract and transform the data to standardize information across all catalogs, enrich the data with human-in-the-loop annotation pipelines, and build quality assurance tools and processes in order to ensure high quality delivery of data.
Given the specification and deadline, it did not seem possible to meet the customer goals. But, after some long nights of sprinting, problem solving, and teamwork, we exceeded our own and the customer’s expectations – so much so that the proof of concept exploded into a very high volume contract. The two week sprint then became a two month marathon to meet the next customer deadline. We had to productionize all our scrappy processes from the proof of concept in order to scale to more than 1000x the volumes.
The team grew from three engineers to around twenty in a short amount of time. I found myself not only executing on my own engineering work but also onboarding and mentoring countless others. Although my title was still officially “software engineer,” I was working on building out parts of the eCommerce product as a product manager and also became the acting engineering manager for some of the onboarded engineers. I was rapidly growing into more leadership-like roles within the team; I owned large surface areas within the product, and led pods of engineers to drive outcomes within those areas. Later, I ultimately stepped up to take on the responsibility of becoming the single threaded owner who led and owned the success of the entire production contract, which involved leading not only all the engineers on the team but also all cross-functional stakeholders.
After my work on the eCommerce team, I transitioned to becoming the first product manager and engineering manager on Rapid, one of Scale’s fastest growing self-serve products after having been launched only one year prior. Now, I spearhead numerous workstreams to push us towards the overall goal of being the default annotation platform for all machine learning and data use cases. Every day, I find myself hopping from thread to thread across engineering, product, design, finance, marketing, and sales. I also still code from time to time to increase engineering bandwidth on the team so that we can push new features for our customers and patch inevitable bugs as they come up.
The challenges on Rapid are unlike those that I have previously experienced at Scale. Instead of building for a specific annotation use case, we need to build generalizable and scalable product solutions in order to support all data types and annotation use cases. This involves being able to support images, videos, text, audio, and other types of data. We also need to develop the taxonomy and tooling to support all associated annotation types, including classification, transcription, semantic segmentation, and geometries.
Given that customers want high quality data as quickly as possible, we also optimize our work to redefine the data quality and turnaround time frontier. To ensure high quality data annotation, we innovate in areas, such as generalizable worker training infrastructure, and ML-assisted validators to flag annotation issues and edge cases. To make sure the data is annotated as quickly as possible without detracting from quality, we also invest in better tooling, ML-assisted annotation (e.g. auto semantic segmentation), and pipeline improvements to break one larger annotation task into multiple smaller ones so that multiple people can work at the same time (e.g. our video stitching pipeline).
While driving all these product improvements and leading some parts of the technical solution, I also partner with our marketing, sales, and finance teams. I experiment with different growth hacking initiatives, such as hosting our first Kaggle competition and partnering with universities. Furthermore, I work closely with our sales team to make sure that we are growing existing accounts and customer segments as well as targeting untapped market opportunities. For every sale opportunity, I am also looped in on pricing decisions in order to make sure that we drive revenue growth while balancing our costs and margins.
From building upon my raw technical ability to building a diverse suite of skills as a product and engineering manager, I am grateful everyday for the opportunities that Scale has given me to build my career. I will never forget the memories I have made and the skills that I have grown on the Marketplace, eCommerce, and Rapid teams as well as through my heavy involvement in university recruiting and event planning.
The tight-knit community is a constant reminder that we can accomplish great things together while making lifelong memories along the way. Recently, I was somehow able to convince our company leadership (including our very own CEO!) to learn a dance that I taught them for at most thirty minutes. Then, they performed this dance together in front of the entire company during our first company-wide in-person celebration after quarantine.
Scale is a great community with some of the most hardworking, talented, and fun people in the Silicon Valley tech scene. I will always be thankful for the highly talented group of people who were with me through the ups and downs. These past two years have been quite the rollercoaster ride, but my reign as the top Slacker of Scale has been quite consistent.
