I’ve been having a blast since we sold StackStorm to Brocade.
As promised in my last post (the 4Ps of Picking), in this post I’d like to share with you a little bit about what I’m seeing in the market.
Like many, I see a greater gulf than ever between the frontier of what is possible and the performance of the average enterprise. You can see this gulf when you measure metrics in terms of operational agility — such as deployment frequency or deployment lead time (time spent in the queue before production).
And you can see the gap in outcomes in studies such as the Puppet Labs sponsored State of DevOps survey here:
One of my favorite (hello StackStorm and auto-remediation!) is that the MTTR for super high performers is 168x faster than average performers.
So with that as context, here are a few of the spaces into which I’m looking:
Given my background having helped create the open storage and software defined storage space, storage is a natural for me albeit one that is under incredible pressure and stress these days in part thanks to too much venture investment chasing what is a large but mature market. I have a bunch of friends and colleagues in storage and am confident in my picking skills in this space.
I’m really proud that Cloudian has invited me to join their advisory board.
Cloudian has a unique offering in the storage space — it is a massively scaleable 100% compliant on-premise S3 cloud that includes far better metadata than does S3 for use in management (thanks in part to an early bet on AWS and Cassandra). They are achieving great outcomes for their customers both by saving them money on object storage (speeds, feeds and dollars per GB!) and, more importantly, by improving the productivity of IT teams all the way up to and including the development teams.
They have spent literally years perfecting their object storage with customers like NTT hosting and tier one financials. And now the word of mouth is spreading.
Cloudian is a much later stage company than some of the others I advise. It’ll be in the news quite a bit in the weeks and months to come thanks to their well deserved accelerating momentum.
This is the first of a few stealth mode start-ups I am advising. This one is looking at the coming impact of 3D memory. I cannot reveal much other than to say that putting many many TBs of non-volatile memory next to the CPU at nearly memory speeds is insane and wonderful. By the time this company emerges I’m not at all sure it’ll be seen as a storage company; storage as a space has been utterly transformed and yet storage companies are too often in my opinion stuck in the speeds, feeds and $/TB mindset of the early 2000s.
Machine learning and data science:
This is an incredible area to learn about. There is so much hype and yet also it goes without saying that some of this deep learning stuff is getting awfully useful. I am not an expert here, unlike storage, and yet I’ve made it an area of focus in my networking and learning. I’m starting to grok the various camps in machine learning in large part with the help of many of the companies I mention below. I’m also getting hands on with my limited Python chops. Fun stuff.
My sense is that those companies that best focus their AI or machine learning on specific pain points will flourish and that many of the opportunities for platform companies that provide for example “data science as a service” have faded away.
With that in mind I’m extremely excited to be supporting TextIQ. Apoorv, Omar, and the entire team at TextIQ are harnessing cutting edge machine learning to address some real pain points in the legal industry. They have tremendous traction and when you meet Omar and see the demo it is easy to see why — clarity of vision, tremendous energy, high CPU, and yet active listening and more. This a rocket ship on the launch pad; yes — slightly hyperbolic and yet I could not be more bullish on their prospects.
They are hiring — and picking their next handful of proof of concepts and production deployments as well. http://textiq.com
I’m also working with Andy and Xavier at Data Fellas. This team has a track record implementing data science pipelines for some of the larger users in Europe and are leveraging this experience to build related software. They are also prime drivers for the now widely used Spark notebook. You can see Andy’s activity on GitHub here: https://github.com/andypetrella
As the name DataFellas and the tag line “we make offers to data they cannot refuse” both suggest, these guys are fun and a little bit irreverent. More importantly, each time I chat with them I come away more impressed by their understanding of what it is like to deliver an distributed data science pipeline to enterprises. They have spent so much time helping actually drive outcomes for customers that they truly feel their pain.
DataFellas are close to getting their product out in alpha / beta form — and in the meantime are doing workshops with folks doing data science at cost in return for getting additional product feedback. Back in March O’Reilly picked them to do their on-line training “Building Distributed Pipelines for Data Science using Kafka, Spark, and Cassandra” — so their expertise speaks for itself. Get in touch with them now — Andy is speaking in NYC this week and is scheduling chats and at least one training now: http://www.data-fellas.guru
In both cases, as you dig in, you’ll find incredibly energetic teams that have survived rigorous PhD programs and are now doing the real work of building great companies. I’m hugely proud of the progress both teams have already made.
Somewhat in the space as well — although not yet deploying machine learning — is CareerWave. At the highest level CareerWave is sort of like uber for career and business coaching. However it is more than that — we are all told these days to “own our own career.” Ok, but how? Not everyone can afford thousands of dollars a month for a coach and yet study after study suggest that coaching helps lead to happier and more successful people. And maybe more importantly for companies, unhappy people under perform and eventually leave. What if we can apply software and machine learning to the problem? That’s the gold standard of coaching, and it’s CareerWave’s approach — they are signing up betas now and also coaches.www.careerwave.me
There is yet another company in stealth mode that is looking to leverage machine learning for support related tasks. Stay tuned.
- System X
- System Y
Yes, sorry, these are two stealth stage start-ups. Each of them intends to help enterprises better measure and automate their operations — and so at a high level they may seem like the Nth monitoring or orchestration or continuous delivery solution. And yet, each are different in part by explicitly focusing on enterprise adoption as opposed to primarily on community usage. The DNA of these companies, much like StackStorm, merges deep DevOps experience with company building and enterprise operations experience as well.
The founders of both of these systems are already gathering around them an incredible team and some great early adopters as well. I’m bullish on both. And I will share more as their founding teams are ready for me to do so.
I’m now spending about half of my time meeting new companies, attending meet-ups and so forth. The other half is split between helping existing companies and doing some hacking and preparing for various Spartan races.
A couple of other machine learning related companies I’ve just gotten to know are again characterized by brilliant technical teams that are drilling into specific pain points. I think the entire team at Alchemist Acceleratordeserve a lot of credit for helping these teams iterate towards product / market fit quickly — while also shared a lot of otherwise very hard earned knowledge about company building. While I’ve just gotten to know these companies, I think they are both interesting:
- DataCulture — here Karthik and team have drilled into a specific pain point in ecommerce that they are addressing with AI powered software and services. They are about 5 days away from revealing their MVP here:http://supply.ai
- Relato.io — Russell is a well known agile data scientist — after all he wrote the O’Reily book of that name — with a track record building out such capabilities at LinkedIn and elsewhere. He’s now applying and extending his capabilities in order to drive waste out of the sales and business development processes. http://www.relato.io/index.html
If you are interested in someone like me helping you out — or at least hearing you out — please do get in touch. My network of friends and of people that seem to trust me has expanded quite a bit over the years. I’m looking for founders and later stage companies that could use my particular insights, relationships, and drive.
Speaking of drive, one thing I’ve learned since selling StackStorm is that I’m definitely not done yet. I’m having a blast and feel every bit as competitive as I ever have.
Last and likely least I’m also shooting for some upcoming hackathons to test and stretch my Python skills. Here I’m most interested in apps that help support community engagement and that shine a light on our governments. The deeper I dig into my local government the more I see the need for transparency and innovation.
I’m also proudly volunteering time as a member of the Vestry at St Matthew’s Episcopal in San Mateo. It is a warm and welcoming community with inspirational leadership.
My ask for you is — what am I missing? What do you think about my areas of focus? If you were me, what would you do differently?
Please keep in touch.