Siddharth Venkatesh (Product Management Lead at Anyscale)

Sid discusses building open core products, developing user empathy, shares advice for young professionals, and more!

Apr 24, 2023

If you are interested in reading more interviews, just like this one, with founders, operators, and investors from today’s leading companies / startups, feel free to subscribe below!

Guest Profile:

Interview Guest: Siddharth Venkatesh (LinkedIn)

Role: Product Management Lead at Anyscale

Previous: Prior to joining Anyscale, Sid was a Product Manager at Meta, where he managed Ranking & Relevance for the Facebook NewsFeed. Before Meta, Sid was an early PM at Rubrik, where he worked on launching multiple zero-to-one products, including their first SaaS application, Polaris. Sid has held various other positions in product management, engineering, and venture capital.

Quick Note: This interview was recorded via a Zoom call between Sid and Roshan (that’s me) in March.

Roshan’s favorite quotes from the interview:

On building customer empathy: “As a PM, I believe it's crucial for as many people as possible within the organization to hear from customers directly. It's one thing for a PM to say that a customer has challenges or friction with a particular workflow, but it's another thing altogether to be on the call and hear that frustration and pain point from the customer directly.”
On important traits of PMs: “Don't be married to your ideas. It's great to have a hypothesis-driven approach, but you need to know when to change course if something isn't working out. Great PMs are informed by data, not attached to their ideas, and they blend qualitative and quantitative inputs when making product decisions.”
On working on an open core product: “I’m amazed and inspired by the passion and involvement from the Ray community. Unlike with “traditional” products that I’ve managed in the past, there is no separate customer feedback / validation stage because we hear from users at every step. Developers passionately advocate for feature requests, and share use cases we hadn’t considered.”
On podcast recommendations (I was thrilled that Sid liked my rec!): “For podcasts, I love Lenny's Podcast, which you recommended to me.”

Roshan: Hey Sid, thanks for coming on. To start, can you tell us a bit about what Anyscale does?

Sid: Absolutely. Thank you for having me on.
So, Anyscale is the company behind Ray open source – the fastest growing open source compute framework for scalable computing. Anyscale was founded three years ago out of the RISELab in Berkeley by Robert Nishihara and Philipp Moritz, along with the director of the lab, Ion Stoica. Interestingly, this is the same lab where Databricks was founded.
In terms of Anyscale’s mission, Anyscale is on a very ambitious mission. Our mission is to make sure that distributed computing is simple and flexible. We want to ensure that every developer and team can successfully build AI/ML applications without worrying about building or managing the underlying infrastructure. Said differently, we want developers and teams to focus on applications and not on the underlying infrastructure. To make this possible, we aim to remove distributed system expertise from actually realizing this value for each of these teams.
Going into more detail, as I said before, Ray is a unified compute framework for scaling AI/ML workloads. Ray essentially enables ML engineers and developers to scale their applications from their laptop to the largest of clusters with minimal code changes. Ray is Python native and integrates with the entire machine learning and data ecosystem. It provides a set of primitives that allow developers to parallelize their existing Python applications easily. Ray integrates with the entire machine learning and data ecosystem, including all of the different learning frameworks, all of the ML traditional libraries such as ScikitLearn, all of the different workflow, orchestration platforms, experimental tracking, and more. Ray essentially provides a set of primitives that allow developers to parallelize their existing Python applications easily.
Ray by itself has a flourishing community that includes leading technology companies such as Uber, Spotify, Shopify, ByteDance, Instacart, and more. Anyscale is a managed solution on top of Ray, enabling teams to accelerate the development, experimentation, and deployment of AI models from dev to staging to production without managing the underlying infrastructure. Anyscale is like Confluent for Apache Kafka or Databricks for Spark – a managed solution for the Ray compute platform.

Roshan: Can you discuss your background prior to joining Anyscale and how you came to join any scale?

Sid: So I started my career like a lot of other folks in the field, as an engineer. My first experiences were in traditional hardware companies such as Intel, Marvell Semiconductors, and Qualcomm. I worked on building wireless algorithms & optimizing performance for their chipsets. However, I eventually wanted to move into product management, so I made a brief stop in business school at the University of Chicago. After that, I worked at Success Factors, a human capital management product that was acquired by SAP. I then worked at Rubrik, a cloud data management company, where I was one of the earliest PMs. I worked on building multiple zero-to-one products, including their first SaaS application called Polaris. I know you were part of that team as well, and that’s where we met.
After spending around three years at Rubrik, I moved to Meta, where my role was to lead ranking and relevance for the NewsFeed. Essentially, my team was responsible for ranking content from the thousands of posts and pieces of content that we could potentially show somebody when they logged into Facebook. We had to ensure that content was interesting for them, and that it was personalized based on what the users wanted. I found out about Anyscale through some friends and was drawn to their mission and the challenging problems they were trying to solve, some of which I had seen firsthand. I was very impressed by the people and the cutting-edge work they were doing in AI and ML. I felt that if these problems are solved well, they have an outsized impact on the broader set of customers we are trying to serve. So for all of these reasons, I joined Anyscale ten months ago and I'm now a product management lead on the product team here.

Roshan: Let's double-click into what made you interested in product management. What was your path to becoming a product manager and what made you excited about product management?

Sid: People with engineering backgrounds or other professional experiences often have intriguing stories about their transition into product management. I genuinely enjoyed being an engineer. I loved coding and tackling various challenges. However, I was eager to play a more significant role in determining the product's strategy. I wanted to understand the rationale behind many decisions, such as how customers used our product, why we should build it in a certain way, and how to determine pricing. I found myself increasingly drawn to these types of questions.
As an engineer, my primary responsibility was delivering algorithms or code, but that role didn't provide much insight into these areas that excited me. I began participating in more customer conversations, and I discovered that I was more attracted to understanding the problems customers faced. I focused on how our technology could help solve those issues rather than concentrating on specific algorithms or coding aspects of delivering a solution.
In short, I became more passionate about understanding & addressing customer problems and enabling their success, more than the specific implementation mechanisms. Additionally, I realized that I enjoyed working with diverse stakeholders, such as legal, marketing, and sales teams. As an engineer, my opportunities to collaborate with these groups were limited. The combination of engaging with customers, identifying their problems, and working with cross-functional teams ultimately fueled my enthusiasm for product management.

Roshan: Going back to Anyscale, what exactly does your role as a product management lead at Anyscale look like?

Sid: As a Product Manager, I mainly focus on three key values: enhancing developer productivity, reducing TCO and data assurance.
So specifically, as a Product Management Lead at Anyscale, I concentrate on three primary responsibilities:
Developing new business charters and go-to-market capabilities to enable the key value drivers
Creating delightful zero-to-one experiences in the areas of data security, cost management, network & access controls, and governance
Hiring great people and fostering high-performance teams.

Roshan: How do you think about building a great product culture at Anyscale?

Sid: At Anyscale, we have an extremely customer-focused culture. The customer is at the center of everything we do. Given our stage, there's a strong sense of urgency, and the company is very scrappy. Our product culture is dedicated to moving fast, so speed is critical. As a product organization, we plan, drive alignment, and execute quickly. Anyscale also has a humble and low-entitlement culture, with talented individuals who are helpful and possess a strong team mindset.
This approach leads to building high-performance teams in three ways. First, we develop a sense of psychological safety within the company. When moving fast, there's always a chance that a bet we make might not succeed. That's okay; we learn as we go. We create psychological safety so that people don't become risk-averse due to the fear of making mistakes.
Second, we continue to develop empathy, which is a key aspect of any good relationship, whether it's between sales and PM or PM and engineering. Fostering trust and empathy is critical for building high-performance teams.
Lastly, once trust and empathy are established, we work on shortening feedback loops. It's essential for us to learn from each other by providing actionable, specific feedback that helps us improve week by week, month by month.
So, these are the three broad ways I think about creating high-performance teams at Anyscale.

Roshan: That's interesting. Can you explain how you build customer empathy within your team?

Sid: When it comes to customer empathy, I think there are three different things that we've done in the past. First, many organizations have very clear distinctions between roles and responsibilities, where it's only the PM who's responsible for talking to the customer, and it's the role of engineering to build things out. We try to avoid that. As a PM, I believe it's crucial for as many people as possible within the organization to hear from customers directly. It's one thing for a PM to say that a customer has challenges or friction with a particular workflow, but it's another thing altogether to be on the call and hear that frustration and pain point from the customer directly. So, building customer empathy involves engaging a set of people early-on in the customer engagement process.
Second is shortening the feedback loop - we actively engage with customers and since Anyscale is an early-stage company, many of them have taken a bet on us. Having frequent touchpoints, making these customers feel heard, bringing them along on our journey, discussing with them how we're thinking through building features and prioritizing our roadmap, and getting their input helps us with user empathy as well.
Finally, let's say we've heard from a customer and have a set of requirements from them, either for a new feature or an improvement to an existing workflow. We have an amazing design team that is actively involved in design studies & customer validation. Conducting these design studies and observing how customers interact with the product goes a long way for us to understand the different friction points they have
So, these are three ways in which we, as an organization, can help increase user empathy.

Roshan: That's great to hear. I like the idea of getting everyone in the organization involved in customer calls.

Sid: That's a great way to build empathy –– one of the things we also do is hold weekly sessions where our PM, design, and engineering teams watch customer conversations together. During these sessions, we all sit together in a room, watch the calls, take notes, and identify areas for improvement based on customer feedback. This approach helps us maintain a shared understanding of the customer's needs and perspectives, which is essential for developing better products.

Roshan: What do you think is the key to great product and engineering collaboration?

Sid: One of the challenging aspects of being a product manager is finding a way to successfully collaborate with the engineering team. As a PM, you have an abstract business idea or feature in mind, and you're responsible for making it tangible. However, as a PM, you work through others to build it out. Collaboration is crucial and enables teams to do their best work together. I'll discuss three key aspects of effective collaboration.
First, building a culture of trust and empathy is vital for great PM-engineering collaboration. Establish an environment of mutual respect and support where teams can work towards a common goal. Spend time getting to know each team member, their roles, and their underlying motivations. Different engineers have different interests and career stages, so understanding their motivations is crucial. Additionally, providing information about why certain decisions were made helps build trust. It's important to reason out your decisions and communicate them to the team. Finding common interests outside of work can also strengthen trust.
Second, aligning towards a common goal focused on the end user is essential. Effective collaboration requires a shared focus on understanding and addressing the user's needs. Engineers can offer valuable insights on design problems or the feasibility of new workflows, allowing them to voice their opinions and ensure they feel comfortable expressing their ideas.
Lastly, having continual feedback is vital to ensure everyone's aligned on goals, priorities, and key metrics. In a supportive environment, team members can openly communicate when things aren't going well and rely on the PM to have the engineering team's best interests in mind during conversations.

Roshan: As this is the first open-core product you've worked on, how do you think about product management differently at an open-core company compared to the previous companies you've worked at, which weren't open-source?

Sid: This is a very different business model than my previous experiences, and is one of the main reasons I joined Anyscale. I've observed differences in multiple areas:
Firstly, when working on open source, you interact with a passionate community of developers who are experts in the area and may have contributed significantly to making Ray successful. I’m amazed and inspired by the passion and involvement from the Ray community. Unlike with “traditional” products that I’ve managed in the past, there is no separate customer feedback / validation stage because we hear from users at every step. Developers passionately advocate for feature requests, and share use cases we hadn’t considered. By conducting the process in public, it’s higher in volume and frequency than anything I’ve experienced. This is a very different ecosystem compared to a PM who typically interacts with business stakeholders, CFOs, or sales teams.
The go-to-market strategy is different since on the open source side, building a strong community, creating forums for collaboration and sharing knowledge, creating awareness & distribution through developer community channels, evangelism and partnerships to help expand reach and create new opportunities, are critical.
Thirdly, the decision of what features to keep in open source versus which ones to monetize is a balancing act between providing value to the community, building a sustainable business model, and competing in a crowded space. We need to maintain transparency, consistency, and clear communication to ensure our open source community continues to thrive and drive adoption, while still providing significant value to customers who pay for a managed solution.

Roshan: HashiCorp is also open-core so that definitely resonates with me and the work I do. One last question on Anyscale, do you have a favorite feature that Anyscale has shipped or will ship in the near future?

Sid: There are many exciting features, but one I'd like to talk about is called Workspaces. Workspaces is a feature Anyscale has built to help customers accelerate the development and scaling of ML and Python workloads. One key problem we've seen is environment reproducibility. The goal is to help developers iterate faster on models and a common, reproducible environment makes it easy for multiple developers to collaborate, debug errors, and allows new developers to continue existing projects.
Workspaces provides an integrated IDE experience, allowing multiple developers within a team to edit and run code, install dependencies, and monitor jobs, services, and resources just like they would on their laptop. It creates a shared environment for everyone to work together.
Additionally, Workspaces offers integrations with tools that developers are already familiar with, such as VS Code, Jupyter Notebook, GitHub, and Weights & Biases for experiment tracking. This allows developers to continue using the tools that make them efficient.
Lastly, we've made collaboration easy. Workspaces can be shared and cloned with a click of a button, allowing different users to access the same configuration and environment, boosting productivity. These capabilities make developers more productive, increase time to market for organizations, and simplify the entire process.

Roshan: What advice would you give to young professionals looking to go into product management?

Sid: That's a great question, and one I hear quite often. Firstly, it's important to understand that product management is a unique role, and its responsibilities can vary across companies and even within different teams in the same organization. There are misconceptions about product management, like the notion that a PM is like a CEO of a product, which, in my experience, isn't accurate.
Product management requires influencing without authority and achieving success through others. In other words, PMs are never in control of outcomes, which can make the role stressful. However, it can be a fulfilling career for those who enjoy solving business and customer problems and working with diverse people.
To prepare for a product management role, focus on improving the following areas:
Critical thinking: Continuously question whether you're solving the right problem and consider the long-term implications and unintended consequences of decisions.
High agency and proactivity: Develop the ability to find a way to achieve your goals without waiting for ideal conditions, pushing through challenges and making progress.
Trade-off identification and balance: Every decision a PM makes involves trade-offs. It's crucial to think through and articulate these trade-offs, as well as the rationale behind each decision while considering potential downsides.
Focusing on these three areas will be beneficial when transitioning into a product management role.

Roshan: What do you think are the most important traits and skills of product managers?

Sid: The first one is developing deep user empathy. At the end of the day, it's crucial to understand what users are trying to accomplish and what constitutes success for them. Users might say they want one thing but in reality mean something else, so being able to empathize and understand their core needs is essential.
The second trait is cutting through ambiguity and providing clarity. With shifting market dynamics, macroeconomic situations, conflicting priorities, and changing competitive landscape, it's important for a PM to think deeply, reduce ambiguity and provide clarity for other stakeholders.
The third trait is ruthless prioritization. A PM needs to cut through the noise and prioritize effectively. This involves finding the right balance between long-term strategic aspects of the product and responding to immediate customer feedback. PMs should always consider industry trends, product positioning, potential partnerships, competitive analyses, and pricing.
The fourth trait is versatility. Don't be married to your ideas. It's great to have a hypothesis-driven approach, but you need to know when to change course if something isn't working out. Great PMs are informed by data, not attached to their ideas, and they blend qualitative and quantitative inputs when making product decisions.
Additionally, clear communication is vital. PMs work with many different types of stakeholders, so it's important to tailor communications and switch context instantly, depending on the stakeholder, to ensure the message is clear. These are some of the critical traits I believe a PM should possess.

Roshan: Outside of work, what hobbies occupy most of your time? How do you stay physically and mentally fit given the high demands of your job?

Sid: I have a five-year-old daughter who keeps me very busy with activities like swimming and dance classes. So, spending time with her takes up a lot of my time.
In addition, I'm deeply involved in nonprofit activities, which is another passion of mine. I've been engaged with nonprofits for over a decade, holding leadership roles in organizations addressing senior care, financial literacy, and computer education.

Roshan: Do you have any favorite books, movies, podcasts, et cetera, that have been a big influence in your life as a whole or in your work?

Sid: Absolutely, there are some fantastic resources out there, especially now that product management has become a more mature discipline. I'll share some recommendations across various formats since people have different preferences for consuming content.
In terms of books, some of the best ones I've read include "Inspired" by Marty Cagan, which is an excellent book on product management. Another great book is "Continuous Discovery Habits," which provides a guide on gathering requirements and understanding pain points during all phases of product development. It's essential to stay connected to customer experiences as a product matures, and this book offers valuable frameworks to help with that. Lastly, "Good Strategy, Bad Strategy: The Difference and Why It Matters" is a fantastic read on strategy.
For podcasts, I love Lenny's Podcast, which you recommended to me, so thank you for that. Another great one is The Skip Podcast by Nikhil Singal, who was the VP of the organization I was in at Facebook. It's tailored for more senior product management professionals, featuring many chief product officers, but it's an excellent resource overall.
As for blogs, Shreyas Doshi has a fantastic blog and Twitter account with insightful posts. Two other great blogs are by John Cutler, a product leader at Toast, and Gibson Biddle, who was a leader at Netflix.
If your audience is interested in courses, Shreyas also has a Maven course that I haven't taken personally, but I've heard fantastic things about it. It might be useful for some of your readers.

The Takeoff

Discussion about this post