Life as a PM on the Splunk Machine Learning Team

Starting a new job is stressful any time. Starting a job in the thick of a pandemic-enforced shelter-in-place is its own beast. I learned this firsthand when I started a new job in May 2020 with a team that I never met face to face, not even for interviews. Towards the end of 2020, I then got an opportunity to interview with the Machine Learning (ML) Product Management team at Splunk. Even though I remembered the experience from 6 months prior of onboarding and starting a new job remotely, I jumped at the opportunity, and started in January 2021. 

Some Background On Me

Before coming to Splunk, I worked in application security — static application security more specifically. I loved every moment of working on the hard problems of finding vulnerabilities in application source code. It is a very complex problem to solve, especially when done with high accuracy and high performance. It is also very rewarding and I felt like a superhero everyday — solving important problems that affect a lot of people alongside some very brilliant minds. Obviously, that is what I was looking for in my next role, also — challenge, talent, ownership, and responsibility. It has been six weeks since starting my new role at Splunk ML and I wanted to share my experience of starting remotely and my thoughts on our portfolio.

Onboarding was a breeze. In my first week on the Machine Learning PM Team, I went through a bootcamp with other new hires. The focus of this bootcamp was to give us insight into Splunk’s different product lines as well as Splunk's culture. It was super well-organized and fun. At the same time, I started meeting my new team members virtually. Everyone I have met so far at Splunk has been very helpful, welcoming, and nice. More than anything else, this is what I have liked most about my Splunk experience.

After the onboarding and “meet and greets," I experienced what my hiring manager had warned me about — “You will be drinking from the fire hose.” I was and I still am, in a good way. As you will see later on in this post, we are building a lot of cool things in this team, which is challenging but also exciting. The ML team moves fast, is not shy of challenging rules that don’t make sense for our team and our customers, and paves its own path. If something needs to be done, we figure it out and we do it. (If this sounds like you, we are hiring! Check out the many machine learning roles at Splunk)

Which brings me to the very important question of what is it we do here in ML at Splunk —what products are we working on, what problems are we solving, and for whom?

Starting with why, our mission is to empower Splunk customers to leverage machine intelligence in their operations.

What Problems Are We Solving?

Our team’s main goal is to enable customers to develop new advanced analytics & ML workloads on their data in Splunk, thus increasing the value they realize from the platform. We want to increase engagement, enable new use cases, and enrich the Splunk experience for our customers.

Who Are We Solving These Problems For?

We strive to make machine learning accessible to all Splunk users. Currently, our offerings meet the needs of four different personas that range from novice to expert when it comes to familiarity with data science and ML:

1. Splunk app users
2. Data analysts
3. Splunk admins
4. Data scientists
   
   

How Are We Solving These Problems?

The different personas we are serving require different solutions — from no-code experiences to heavy-code experiences. We achieve this by having a breadth of products:

• No-code ML: Allow customers to unlock insights from their data without data science skills. This is available as MLTK (Machine Learning Tool Kit) today. However, we are working on the next generation of this product which will be simpler to use, more scalable and enterprise ready. In the future, it will be available as part of SMLE (Splunk Machine Learning Environment)
• Virtual assistants/ Smart workflows: Provide building blocks to customers to create their own solutions
• Use case blueprints: Ready to use notebooks for security, IT and observability use cases
• SMLE Studio: Jupyter notebook experience for the data scientist and researcher
• SDK: Allow developers to connect to SMLE from their applications via a programmatic interface
  
  

Where Will Our Solutions Live?

These products cover the different personas we are targeting. However, we need to make it easy for users to use our solutions where they are. We achieve this via the following:

• A central console or location for content management, 1-click deployment, and monitoring across the ML product family
• Pre-packaged content including reference solutions, blueprints, and smart workflows apps accessible from the SMLE Console
• Ready-to-use operators like drift detection and automatic field extraction embedded in our platform services offerings
• Embedded ML-powered features like smarter alert management, advanced threat detection in our IT, Security, and Observability premium apps
  
  

It is evident that we have a bold vision and lots to do. We want to make ML-powered insights accessible to core Splunk users. At the same time, we want data scientists to be able to leverage their Splunk data within Splunk.

In the past one year and for the short term, our focus is on the data scientist. We are working on making SMLE Studio available as an app on Splunk Cloud Platform. However, for the middle term, we are going to shift our focus to the Splunk user. 

There are other initiatives in Applied ML and research, streaming ML, and the embedded ML space. I will leave that for another blog post because, as I said earlier, I am new! I’m still learning, and there’s so much to cover!  

The most exciting part for me is that we are in the early stages of delivering on this vision. There is a huge opportunity to own a big part of this effort and create an impact. Ask any product manager and you will quickly know that more exciting words have never been spoken. Needless to say, I am very excited about all the amazing things we are going to build together. Onwards!

Want to help us tackle this vision? Take a look at our machine learning roles today.

Resources

• Interested in SMLE? Sign up for our Customer Advisory Board and interest list  
• Machine Learning Guide: Choosing the Right Workflow
• Using SMLE and Streaming ML to Detect Anomalies
• SMLE Product Brief
• We are hiring! Join the team.