5 Major Lessons Learned as an ML Engineer Turned Head of Dev Rel & Solutions For an Open-Core MLOps Startup. Part 1/2.
Or How I Try To Sneakily Explain Where I've Been For The Past 6 Months.
🕰️ First, Let’s Rewind To Last Year 📅
Hey everyone!
So it’s been a long, long time (at least in social media terms) since I’ve posted on Substack.
My last update was posted approx. 6 months ago, around the time I started my new role at Featureform.
Although I didn’t COMPLETELY stop writing, speaking, or sharing content about MLOps, I did have to scale back for a while, especially as super exciting things were happening with Featureform.
Consider this a general update on what’s been going on & where I’ll be taking this Substack.
My TL;DR:
I’ve been learning a lot about myself as an engineer, content creator, and tinkerer. One of those learnings is that I do really badly trying to juggle many activities & goals so for the foreseeable future I’ll be limiting myself to 5-6 main activities this year.
Specifically I’ll be focused on writing & publishing on Substack, creating videos for Youtube, and creating projects on Github.
What I will NOT be doing: Investing serious time & energy into TikTok, Instagram, other publishing platforms, streaming off Twitch or streaming in general. You will also not see me entering the podcast realm and if I do either podcasting or streaming it’s going to be hosting a panel of much smarter people.
I will also not be writing any books this year but I may produce courses closer to the end of the year.
With that being said, I’m not making any promises on frequency of posting. My goal is to write that which can’t be easily ChatGPT’d and that takes time, effort, and doing my own learning and coding.
This post series is going to be a bit in the style of stream-of-consciousness so bear with me as I’ll share some learnings that are super important if you’re thinking of:
Jumping in your first DevRel role;
Jumping into full-time content creation;
Diversifying your content & revenue streams;
Jumping to an Open-Core company.
Lesson #1: DevRel ≠ Just Content & Solutions ≠ Just Implementations
🖊️ Main Points:
DevRel (or Developer Relations) encompasses many different strategies and tactics for getting products & projects in the hand of users & then taking the feedback to product & engineering to help the product reach better product-market fit alignment.
Startups can take months & years to reach product-market fit & most importantly product-market fit is a moving target.
While many DevRels are actually practicing Developer Evangelism via Content Creation, there are many examples of DevRels (especially in the early stages of a project) leveraging non-content channels & strategies to develop integrations, contribute to the open-source project, grow communities, etc.
Ultimately the point of DevRel is to increase usage and adoption of projects, oftentimes by increasing top-of-funnel brand awareness and/or deepening usage.
Similar with Solutions, it’s not always about just technically implementing the product or project into the customers hands and then disappearing, effectively tossing the product over to the user/customer side of the fence.
Sometimes there is a really strong need for education, requirements gathering, alignment, & even at times saying “Hey, we don’t seem to be a great fit right now, so let’s pick up the conversation in the future” or even “We don’t do that right now but maybe we can partner and help each other grow”.
🔥 My Spicy Take 🌶️
In short, if you go into DevRel thinking you’re just going to be working on snazzy YT vids and livestreaming on Twitch (aka subsidized content creation), you’re probably not wrong but you’re definitely not right.
This was a learning and realization that crystallized over many, many months.
Although I’d been talking to the fine folks (Simba & Shab) at Featureform on-&-off up to that announcement on Linkedin (& Substack), I didn’t really understand what my role would entail or even where my personal content plans would fit in to the day-to-day.
Every early-stage startup is different.
Before joining Featureform I had experience with two early stage startups (less than or equal to 10 employees) with at least one startup being pre-product.
And a lesson that I conveniently forgot over and over again was that a large part of being successful in the startup game is being flexible with who you are and your scope of work.
Aka pivotting.
Going in I had a wobbly expectation that a good chunk of my work would be focused on content creation, partnerships, and fixing some docs here and there.
So basically, as swyx would call my imagined role in “Measuring Developer Relations”, “very expensive affiliate marketing”.
I began the journey of educating myself on what DevRel was, including following the thought leaders of Twitter Dev Rel, joining a Devs Who Write Retreat in Miami (which. was. AWESOME!), & of course picking up copies of Lewko & Parton’s “Developer Relations”.
Also started a Discord for Dev Rels working on OSS, Open-Core Data Science & ML Devtools called “The Data & ML DevRels Guild” that’s been really nice for getting feedback on questions like “How are you being measured?”.
However in the past couple months we’ve seen massive layoffs across Tech and some of these layoffs have also hit DevRels and Solution Consultants.
And while I believe DevRels & SC’s can be best positioned to survive layoffs & recessions (after all, some have been incredibly successful in building personal brands that include sponsorship deals, ad revenue from platforms like Youtube, and even cohort-based mentoring services), I don’t think everyone is going to have as easy a time getting a new role.
And the reason is as simple as….
Lesson #2: You’re Either Growing Product, Revenue, or Users or You’re Deadweight
🖊️ Main Points:
DevRel, Growth, Product, Solutions… titles largely don’t matter, especially in the early-stages. Outcome matters.
And regardless of the metric, the outcome for a startup needs to be growth in product, users, or revenue.
Why?
Revenue: It’s a business, at the end of the day. Unless an organization is explicitly an NGO, a non-profit, or even a government organization, the organization is in the business of making money. And if a business is investor-backed, the investment was made in the good faith that the company will eventually return revenue (and eventually profit).
Users: This outcome is interesting and there are a number of reasons why users (both free & paid) matter. More users leads to more product feedback, which makes the product better, which leads to more users (hopefully). Most importantly better product means more revenue.
Product: Investments in product should lead to an increase in revenue and users (which leads to an increase in revenue). And assuming that all products are created with the goal of solving a unique pain-point, better product means better solve for the pain-point.
🔥 My Spicy Take 🌶️
You need to be a growth enabler and most importantly, you need to understand the business of growth.
If all you do is content creation without measuring outcomes and activities, you’re limiting the incredible amount of learning & experiences you could be gaining from being a foundational member of the startup, especially in gaining the first 1K users.
I recently joined Reforge’s Spring 2023 Live Growth Series cohort.
So far it’s been a great experience as it’s covered Growth oriented topics & discussion that I don’t often hear in either the DevRel or Content Creator circles.
I’ve long believed that part of being a great technical partner (whether as a data scientist, MLOps engineer, data engineer, etc) is understanding the “revenue-center” of your company.
Whenever I’ve taken on a new role (especially one that was business-facing) as a practitioner I’ve upskilled my knowledge through courses, video lectures and books.
As a data analyst partnered with the sales and revenue ops team for WalkMe and as a financial analyst for Sunrun I picked up books like “Cracking The Sales Management Code”, “Game The Plan”, “The Challenger Sale”, Porter’s “Competitive Strategy” and “Competitive Advantage”.
When I worked on the real-estate tech startup I geared myself with tools to understand the residential real-estate market like “Your First Year in Real Estate”, “Listing Boss”, and “The Millionaire Real Estate Agent”.
Regardless of the industry the north star is always some form of Money + Customers. The metrics can get real fancy sounding but ultimately they’re variations on Money + Customers.
And the closer your work is to improving these metrics and the more ownership you have over said work, the better the upside.
And if you’re a technical practitioner that understands this? That’s leverage my friends.
For a very, very, unscientific and rough analysis of “proportion of engineering or technical practitioners that understand & are doing growth work deeply focused on early user acquisition & engagement for early-stage projects or startups focused on ML/data devtools” (I know, incredibly specific), here are some anecdotes:
On the first kickoff meeting for this Reforge cohort, there were a series of Slido polls for questions like “What’s your business model (B2C, B2B, Other)?”, “What’s your type of product?”, “What’s your biggest challenge?”, etc, etc.
Out of a cohort of 600+ individuals, all presumably interested in learning about Growth, there were maybe a handful that indicated they were in the devtool space for Data and ML via the poll.
And of the ones that introduced themselves in the community forum, I recognized all the individuals or their startups (although the same might not have been true). Less than 6 were from the Data + ML space of the 40+ intros.
When interviewing for either Solutions, DevRel or even Platform Eng roles in MLOps, mutual friends of mine were regularly contacted for interviews for the same role I was interviewing for. Oftentimes we individually ended up passing on the roles after comparing notes or even independently.