Inference: AI business podcast by Silo AI
By Ville Hulkko
In a world in which everyone is talking about AI, a real track record is hard-earned. Inference dives deep into data driven organizations and their machine learning operations to understand what it means to work with AI. Our guests range from lead data scientists to founders, strategists and engineers. Learn more about what production ready AI means today with Inference by Silo AI. Your host is Ville Hulkko, Co-founder of Silo AI.
S2 E07: Lina Weichbrodt: AI Post Mortem
In this episode, Lina Weichbrodt discusses AI Post Mortems and shares her favorite stories regarding machine learning damage control. Lina Weichbrodt is an experienced ML engineer. Having built her career at Zalando and Deutsche Kreditbank, she has developed a specialty for debugging and understanding performance risks of large-scale, deployed systems. Catch Lina’s recap on debugging near-invisible edge cases, customer preference inclusion, and her takeaways for ML debugging and monitoring ops. #AIforPeople #InferencePodcast
September 19, 2022
S2 E06: Michael Munn, Google: Advanced Solutions Lab
Meet Michael Munn from Google, discussing the Advanced Solutions Lab. In the latest episode of Inference, Ville Hulkko of Silo AI sits down with engineering specialist of ML design pattern and explainable AI, Michael Munn. Michael walks us through the operations behind Google’s Advanced Solutions Lab. Catch Michael’s insights on the future of organizational ML adoption, and learn about the evolution behind AI education and the democratization of AI in the industrial space.
August 16, 2022
S2 E05: Errol Koolmeister, The AI Framework: AI strategic advisory
S2 E5: Errol Koolmeister, CEO of The AI Framework, talks about the role of strategic AI advisory. Errol is one of the most well-known figures in the Nordic AI scene. Known for his work as the AI lead at H&M, Vodafone, and Nordea, Errol has developed his expertise in pioneering practical large-scale AI implementations. Recently, Errol has “gone rogue”, helping companies make intelligent AI management decisions. In this episode, we learn about the managerial challenges of real-world AI implementations and dive into the recurring themes of AI problem-solving. Catch Errol's take on AI in talent acquisition, infrastructure, and the optimal AI management models, as well as his predictions on the upcoming trends of AI management.
June 29, 2022
S2 E4: Douglas Castro, NeuralDSP: AI changing an industry
Douglas Castro, NeuralDSP: AI transforming an industry. In S2 E4 of Inference we deep dive into audio processing as NeuralDSP’s CEO Douglas Castro discusses how Machine Learning has transformed an entire industry’s approach to technology. NeuralDSP is the market leader in professional grade audio processing. Their guitar- and bass amplification modeling software, as well as Quad Cortex computation device, have been adopted by both industry professionals as well as advanced hobbyists globally. Their approach to combine DSP with Machine Learning has enabled the creation of extreme efficiency and high quality products. Doug is a known pioneer in the industry. Prior to foundation of NeuralDSP, his previous company, Darkglass Electronics, created the de facto bass amplification system of modern rock and metal music, used by artists such as Faith No More, Periphery and Foo Fighters. Catch Doug’s description how a scientifically extremely complex processes could be scaled with Machine Learning technologies, how the research adoption rate of Machine Learning in audio processing, and learn his predictions regarding the future trends of AI in professional audio. #AIforPeople #InferencePodcast
June 13, 2022
S2 E3: Prof. Timo Vuori, Aalto University: AI as a manager
Timo is a professor at Aalto University ’s Department of Industrial Engineering and Management. Timo is also the co-author of Platform Strategy: a recent book discussing business transformations enabled by AI and platform business models. With Timo, we discuss about the implications that AI will have in management, peopleops, talent matching and leadership. Tune in to catch Timo’s insights into the “whats and hows” of an AI-as-a-manager organization, learn about the key challenges of skill classification and learn Timo’s predictions on the future developments of AI in management processes.
May 23, 2022
S2 E2: Lina Weichbrodt, DKB: AI diagnostics, usability and post mortems
S2 E2: Lina Weichbrodt of Deutsche Kreditbank discusses AI diagnostics, usability and post mortems. Lina Weichbrodt is the Machine Learning Lead Engineer at Deutsche Kreditbank, formerly at Zalando. She has pioneered Machine Learning -focused quality management, debugging and post mortem practices throughout her career and shares her insights regarding ML quality & problem discovery. Machine Learning quality monitoring can be notoriously difficult, due to an absence of the “it either works or doesn’t” -logic common in traditional software applications. Tune in to catch Lina’s recommendations on which PeopleOps principles have a strong indirect affect on algorithmic quality, how ML quality ownership should be allocated, and tips for the best toolkits & recommendations for product owners on how ML quality should be approached. #aiforpeople #inferencepodcast #siloai
May 09, 2022
S2 E1: Daniel Wroblewski, EQT: Motherbrain
NEW SEASON: Episode 1 - Motherbrain, Daniel Wroblewski, EQT Profile photo colorized by a Generative Adverserial Network. EQT is one of the leading private capital powerhouses in Europe, investing heavily in AI and other cutting-edge technologies. To elevate their investment ops, EQT created the “Motherbrain” data & AI platform. Daniel Wroblewski, EQT:s former Chief Architect and Chief of Staff and the head of Motherbrain, walks us through the “why:s and how:s” of creating a private capital AI platform from scratch. Since the recording of the episode, Daniel has moved to operate as Managing Director of Alpha Generation Lab of CPP Investments. Tune in to catch Daniel’s insights into the importance of human-centric design, ways to overcome early adoption challenges with a win-win Active Learning loop, and catch his predictions on the upcoming private capital and AI trends. Hosted by Ville Hulkko of Silo AI. https://silo.ai/inference
April 14, 2022
Episode 20: Stephen Alstrup, SupWiz: State of AI in Denmark 2022
Inference #20 out - State of AI in Denmark with Stephen Alstrup, Supwiz With the release of the State of AI in Nordics 2021 report, we’ll be taking a look at the Danish AI ecosystem. Professor Stephen Alstrup is one of the most prominent figures of Danish digital affairs. Stephen is a professor of algorithms at University of Copenhagen, the chair of the national digital advisory as well as the CEO of Supwiz. With Stephen, we’ll be taking a look at the state of AI in Denmark within the private sector, public sector as well as academia. Tune in to catch Stephen’s suggestions for AI policy making, POC:s vs. production level implementations and learn Stephen’s recommendations for next concrete steps to build Denmark into a more advanced digital society. Hosted by Ville Hulkko of Silo AI. https://silo.ai/inference
December 20, 2021
Episode 19: Jouni Luoma: DeepRacer
Jouni is a senior AI engineer at Silo AI and the Nordic champion of the globally renowned DeepRacer competition. An enthusiast of signal processing, cloud + local data management, Jouni jumped into the deep end of Reinforcement Learning and simulation-based training to take on the challenges creating an autonomously piloting system. Tune in to catch Jouni’s shock and recovery during his first physical race, learn about the deeper implications of DeepRacer and catch Jouni’s predictions of the upcoming applications of Digital Twins. https://silo.ai/
November 12, 2021
Episode 18: Tero Ojanperä: Platform Strategy
Tero Ojanperä discusses intelligent platforms & platform strategy. Tero Ojanperä is a co-founder and chair of Silo AI, and a professor of practice at Aalto University. This fall, Tero and his co-author, Professor Timo Vuori, released a book called Platform Strategy: Transform Your Business with AI, Platforms, and Human Intelligence. The book describes seven steps to converting a business into a platform. Tune in to catch Tero’s thoughts on upcoming transformation trends caused by platform thinking, understand AI’s role within platforms, and learn from multiple practical case studies on real-life platform strategies. https://intelligentplatforms.ai/ https://silo.ai/
October 27, 2021
Episode 17: Michael Munn, Google: Machine Learning Design Patterns
Inference #17: ML Design Patterns with Michael Munn from Google Information can be lost in translation. As new technologies require a unified framework for discussion, they too require it for semantics. Michael Munn, along with his co-authors Valliappa Lakshmanan and Sara Robinson, released a book called ML Design Patterns to help codify common modeling- and engineering problems, solutions, and approaches into uniform language, aiming to democratize ML comprehension. Michael is a professor and mathematician by background. At Google, he works with Google Cloud Platform’s customer-facing projects and is one of the driving forces behind Google’s Advanced Solutions Lab. Tune in to learn about the difference between academic vs. democratic ML, best practice sharing between Google’s teams and matching of Google’s MLOps principles to customer ways of working, and upcoming MLOps trends. https://youtu.be/hNtJJ5R_T9s
October 05, 2021
Episode 16: Regulation, ethics & geopolitics of AI - Mark Caine, World Economic Forum
Regulation, ethics & geopolitics of AI. AI development advances with staggering speeds across the globe, while the democratization of tools grow the number of developers in every corner of the planet. "Through the explosion of smart devices, more people arguably have access to Artificial Intelligence than clean water today", says Mark Caine from World Economic Forum. In a world where transformative technologies outrun human procedure, defining regulative and ethical boundaries can be an overwhelming challenges. World Economic Forum, the Switzerland-based NGO, has made strong advances in researching the implications of AI politics, as well as spreading the best practices of AI implementation and regulation to the public and private sectors alike. Mark Caine is the head of World Economic Forum’s Global AI Action Alliance (GAIA). In this episode, we discuss the state of AI development across the globe, the different approaches to regulating AI as well as the state of AI ehtics discussion. Mark is a diplomat by background with a keen interest for emerging markets. At WEF, Mark leads government affairs at the Center of Fourth Industrial Revolution alongside GAIA. A former Research Fellow at London School of Economics, his work with sustainable energy policy within African Union Commission has gained global attention. Mark is also a known lecturer at LSE, Stanford, UCL, and more. Tune in to catch Mark’s insights on emergent AI hubs outside of “the big three”, the paradigms of AI regulation vs. AI-inclusive regulation, and catch Mark’s predictions on the future development of AI politics. https://youtu.be/bYh5J8kwGA0 https://silo.ai/
September 17, 2021
Episode 15: Roadmap to first AI victories - Anna Mossberg
Anna Mossberg – Managing Director, Silo AI Sweden, ex-Google and Telia Choosing the first AI path of an organization can be a daunting task. Which challenge to tackle first? Where to get the skills and tools to do it? How to manage and budget the process? Anna Mossberg walks us through the steps that organizations both lean and large need to take to successfully embark on their AI journey and implement data strategies. Anna is a technology leadership professional, with a solid background in companies like Google and Telia. Throughout her career, Anna has been the driving force behind data strategies of multiple listed companies, and is a strong advocate of rapid technological experimentation. Tune in to hear Anna’s view on first selecting the first AI path, the role of data strategy, and what different organizational methods exist for early AI adoption. We’ll also get to hear Anna’s concrete three-step program for taking AI from talking to action.
June 07, 2021
Episode 14: Experimentation culture of Netflix - Ville Tuulos
Ville Tuulos - Manager, ML Infrastructure at Netflix, author, entrepreneur Machine Learning is defined by a need for rapid experimentation. To achieve an environment of fast, iterative and low-risk experimentation, both hard aspects (tools and platforms) and soft aspects (culture, ways of working) of the ecosystem need to be aligned. Ville Tuulos, the driving force behind Netflix's Metaflow platform, explains how Netflix has managed to tackle both sides of the coin to build a truly experimentation-oriented organization. Ville is among the most renowned of Finnish data scientists. Spinning out of self-organising map research in Finland, he quickly landed in Silicon Valley at the first dawn of AI market emergence, focusing on the development and implemenation of deployment infrastructures. This led him to push platform development at Netflix, which later became Metaflow, Netflix's Open Source model design and deployment framework. Today, he's spinning out a development platform company, and will be releasing his book, Effective Data Science Infrastructure, this spring. Tune in to hear Ville's point of view on the how a culture of freedom and responsibility is created, the role of processes (or lack thereof), and how the bridge between a thousands of great ideas for ML application should be validated and proven. Get Ville's predictions on the future trends of ML management and human-machine collaboration.
April 08, 2021
Episode 13: Data monitoring & quality - Patrik Tran, CEO of Validio
Data monitoring & quality - Patrik Tran, CEO of Validio “We have good quality data” is a phrase spoken by most organizations. Theory and practice differ, however, as Validio approaches data from a different perspective: all data is flawed by nature, and thus our focus should be on monitoring, understanding and tolerating it. Patrik Tran, PhD, is the co-founder and CEO of Validio, working on a toolkit for data monitoring. With a stellar background in academia and upper management data consultancy, Patrik evangelises a key message: Data is never perfect. Tune in to learn about the famously mismanaged approach to data quality, difference between academic and practical data quality, how data engineers should be empowered within organizations, and hear Patrik’s three predictions on the future of data monitoring within the next years.
March 19, 2021
Episode 12: Predictive Maintenance with HITL - Patrik Strand, Wärtsilä
Predictive maintenance with a human in the loop - Patrik Strand, General Manager, Product Management – Performance Services at Wärtsilä Predictive maintenance is a concept everyone talks about, but few are actually applying. Wärtsilä, an industry leading martitime & energy sector technology and power source company, executes a data-driven strategy through their "Expert Insight" predictive maintenance product. With a strong focus on matching data insights with the deep knowledge base of their engineers, Wärstilä chose a human-in-the-loop driven approach to direct the learning of Expert Insight. Patrik is the general manager of performance service product management at Wärtsilä. With an engineering background, Patrik oversaw the augmenting of their pre-existing Condition Based Maintenance -ops with predictivity, and has a clear vision of the future of predictive maintenance solutions. Tune in to discover how Patrik views the importance of customer collaboration, how interactivity can be translated to annotations, and what Patrik's vision for the future of predictive maintenance solutions is.
February 09, 2021
Episode 11: Internal AI Center-of-Excellence - Girish Agarwal, Husqvarna
Internal AI Center-of-Excellence - Girish Agarwal, Director AI Lab at Husqvarna Among the most interesting management frameworks of AI development is the emergence on internal AI Centers-of-Excellence. Husqvarna is a Swedish global industry leader of power products and robotics, with a product portfolio ranging from forestry and landscaping consumer equipment to industrial scale solutions. With autonomous operations being a cornerstone of Husqvarna’s strategy, the AI lab was established to enable the creation of such systems. Girish is the director of AI Lab at Husqvarna. Tasked with the goal of figuring out for the big data masses collected throughout Husqvarna's operations, Girish oversaw the establishment and management of AI resources and working models within Husqvarna. As a result, Husqvarna is now running one of the most advanced - and successful - internal AI COE:s in the Nordics. Tune in to discover what Girish views as the golden ratio of early AI projects initiated vs. deployed with their ASD Scoring, how Husqvarna's AI management has matured over the years and how exploratory vs. tangible AI initiatives should be approached.
January 13, 2021
Episode 10: State of AI in Sweden - Prof. Hedvig Kjellström, KTH & Silo AI
State of AI in Sweden - Hedvig Kjellström, Professor at KTH, Lead AI Scientist at Silo AI Sweden KTH Royal Institute of Technology’s Division of Robotics, Perception and Learning is among the most renowned AI research organisations globally. Hailing from Stockholm, Sweden, we are joined by Professor Hedvig Kjellström to learn more about the state of research and development of AI in Sweden. Hedvig's pioneering research on artificial sensing is known to most on the industry. In 2010 she received the Koenderink Prize for fundamental contributions in Computer Vision for her article on human motion reconstruction, together with Black and Fleet. Today' she splits her time between academia and working as the Lead AI Scientist in Silo AI Sweden. Tune in to learn about how academia and private sector collaboration has evolved in Sweden, what the hot zones of Swedish AI development are. Don't miss Hedvig's prediction on the next trends of AI research and implementation.
December 19, 2020
Episode 9: State of AI in Finland - Alexander Törnroth, Head of Finnish AI Accelerator (FAIA)
State of AI in Finland - Alexander Törnroth, Head of FAIA Finnish AI Accelerator is the national support entity for Artificial Intelligence, accelerating Finland’s adoption of AI in both the private and public sectors. In October 2020, FAIA released a report titled “State of AI in Finland 2020”. The report, first of its kind, outlines the development of AI adoption in Finland, reviews key enables and obstacles of AI onboarding, and outlines winning case studies of AI implemented in Finland. Alex is known as an influential figure in the Finnish startup scene and is known for his work as the former CEO of SHIFT business festival, as well as the head of startup services at KPMG. He shares his time between FAIA and Silo AI where he heads ecosystem and partnerships. In this episode, we learn the biggest blocks of AI adoption across industries, learn Alex's cherry-picked case studies that represent exceptional AI success stories, and get Alex's prediction on the development of AI in Finland over the next 24 months.
November 26, 2020
Episode 8: AI Toolkits - Luka Crnkovic-Friis, CEO of Peltarion
AI Toolkits - Luka Crnkovic-Friis, CEO of Peltarion Peltarion is one of the leading star companies of Nordic AI, building an operational AI platform that seeks to make AI accessible to not just AI scientists, but to a much broader user base. Hailing from Stockholm, Sweden, Peltarion's platform has already been onboarded by organisations such as NASA, Tesla, General Motors and many more. Luka is a prolific AI scientist who, together with his co-founder Måns Erlandson, pioneered an early version of the Peltarion platform already in 2004. Luka shares his predictions of the future of how AI will build AI, as well as share his favourite examples of how AI-driven data augmentation has led to winning results. Tune in to discover how both Peltarion and Silo AI aim to keep the AI developer talent in the Nordics, and make sure Nordic companies have access to the best AI expertise and tooling.
September 23, 2020
Episode 7: Media & AI - Jaakko Lempinen, Head of CX of YLE
Jaakko Lempinen, Head of CX of YLE. YLE, the Finnish National Broadcasting Company, is a well known industry leader in digital media experiments and technological innovations. As such, YLE has been among the first in Europe to begin implementing AI on a broad spectrum across their operations and content types. Jaakko is a pioneer of designing and leading AI initiatives across content types and -platforms. YLE Areena, a leading public streaming platform, is YLE’s stomping ground for experimental AI initiatives. Tune in to get Jaakko’s lessons learned about successfully executing AI initiatives. Tune in to hear more about ghost stories created by AI with Silo.AI’s & YLE’s generative modeling initiative.
September 09, 2020
Episode 6: Impact, data & AI - Annu Nieminen, CEO of Upright Project
Impact, data & AI - Annu Nieminen, Upright Project In our sixth episode, we dive into the world of impact data. The Upright Project seeks to collect and universalise corporate impact metrics into one, standardised, comprehendible set. Find out how Upright uses AI to extract novel data sources and integrates them into their larger model. Annu Nieminen is the CEO and founder of the Upright Project. Previously is known as the CEO of Kasvuryhmä, a renowned group of scale-ups and established growth companies. With a track record at McKinsey & co, Annu has a pragmatic approach to data and its commercialisation methods. Find out also about Silo.AI's work with climate impact prediction together with Riskthinking.AI.
August 26, 2020
Episode 5: Speech & AI - Otto Söderlund, CEO of Speechly
Speech & AI - Otto Söderlund, Speechly In our fifth episode, we discuss the the rising creed of speech user interfaces and the pivotal role AI technologies play in their emergence. Speechly is a unique speech understanding & Natural Language Understanding technology company, formed around ex-Apple Siri developers. In this episode, we discuss the status of speech technologies globally, and focus on local languages and the strategic and political importance of having independent technology available. Otto Söderlund is the CEO of Speechly. A seasoned technology entrepreneur with an exit under his belt, Otto has been pioneering language understanding development in the private sector. Find out Otto's predictions for the next 24 months of speech UI emergence, and get his take on state of speech tech in Finland vs. in Netherlands, in the UK, or in Sweden.
June 26, 2020
Episode 4: MLOps - Eero Laaksonen, CEO of Valohai
MLOps - Eero Laaksonen, Valohai In our fourth episode, we discuss the emerging creed of MLOps - the art of managing Machine Learning workflows as a part of operations. Valohai is a pioneering MLOps & ML deployment platform startup that focuses on enabling the deployment of complex data science workflows through the Valohai platform. In this episode, we discuss different creeds of ML deployment platforms, and hear Eero's opinion about the current state of MLOps across company types. Eero Laaksonen is the CEO of Valohai. With background in technology innovation consulting and deep tech development, Eero manages Valohai's team consisting of senior level data science experts from both Finland and Silicon Valley.
June 18, 2020
Episode 3: Wearables & AI - Heli Koskimäki, Science team lead at Oura
Oura - Wearables & AI In our third episode, we discuss the Oura Ring - a tiny ring-sized wearable device that gathers data on sleep, activity, readiness, and more. During the episode, we explore how machine learning can be used to create a personalized experience for Oura users, and what is it like to work with highly science-driven teams. Heli also reveals her learnings from working next to the customer service team. Heli Koskimäki is the science team lead of Oura. An elite research scientist from University of Oulu, who branched her data science skills into the wearable industry. Ville Hulkko is Co-Founder at Silo.AI, the largest private AI lab in the Nordics. Prior to Silo.AI he co-founded Valossa Labs, a computer vision startup. Ville lead the company to Silicon Valley, where he also founded Blackbear Startup Incubator.
June 05, 2020
Episode 2: Building AI products & teams - Marc Dillon, CTO of Basemark
Building AI products & teams - the art of combining human, machine and design. We discuss how the perception of AI products has shifted from core technology to holistic design, and map out changes in team structures. Human-machine interfacing is becoming a critical part of almost every digital process - find out what demands it sets for teams. Marc Dillon is a renowned technology entrepreneur and the active CTO of Basemark. Previously known for his work as the Co-founder and figurehead of Jolla, a telephone industry challenger platform, Marc has been a central part of building products and offerings utilising Machine Learning as the centre of core tech stack. Ville Hulkko is Co-Founder at Silo.AI, the largest private AI lab in the Nordics. Prior to Silo.AI he co-founded Valossa Labs, a computer vision startup. Ville lead the company to Silicon Valley, where he also founded Blackbear Startup Incubator.
May 29, 2020
Episode 1: TinyML - Zach Shelby, CEO of Edge Impulse
TinyML - AI pushed to the edge, delivering larger impact in smaller packages We discuss the state of edge technologies, emergence of AI in microcontrollers, use cases of TinyML ranging from maritime to savannah plains, and predict what’s going to happen in the future of edge AI. Zach Shelby is the CEO of Edge Impulse, and a seasoned veteran of edge technologies. Zach sold his previous IoT technology company, Sensinode, to ARM in 2013, he has since directed ARM’s development and edge technology teams. In 2019, he embarked on another entrepreneurial venture to co-found Edge Impulse, a Tiny Machine Learning platform, aiming to democratize AI in smaller packages. Ville Hulkko is Co-Founder at Silo.AI, the largest private AI lab in the Nordics. Prior to Silo.AI he co-founded Valossa Labs, a computer vision startup. Ville lead the company to Silicon Valley, where he also founded Blackbear Startup Incubator.
May 20, 2020