Forward

We sat down with Fiddler’s founder and CEO Krishna Gade to talk about explainable AI, lessons from Microsoft, Twitter, Pinterest, Facebook, Bing vs. Google, raising $45M and what should entrepreneurs building in AI today focus on.

Fiddler is a pioneer in enterprise Model Performance Management. Data Science, MLOps, and LOB teams use Fiddler to monitor, explain, analyze, and improve their solutions and build trust into AI.

The unified environment provides a common language, centralized controls, and actionable insights to operationalize ML/AI with trust. It addresses the unique challenges of building in-house stable and secure MLOps systems at scale.

Unlike observability solutions, Fiddler seamlessly integrates deep XAI and analytics to help you grow into advanced capabilities over time and build a framework for responsible AI practices.

F500 companies use Fiddler across training and production models to accelerate AI time-to-value and scale and increase revenue by connecting predictions to business context or providing visibility into model performance and behavior.

Fiddler is backed by Bloomberg Beta, Amazon Alexa Fund, Insight Partners, Lightspeed Venture Partners, Haystack, and Lux Capital, among others.

0:00 - Will we ever be able to explain AI?

8:33 - Getting a better understanding of why AI system acts the way it does

20:00 - Fiddler’s customer profile & success stories

27:50 - The mechanics of data perturbation

33:20 - R&D breakthroughs powering Fiddler

39:21 - Lessons from Microsoft, Twitter, Pinterest, and Facebook

44:31 - Bing vs Google

47:55 - Founding Fiddler: building team, fundraising

57:20 - Specialized types - a group of ex-Twitter engineers

59:20 - Advice to entrepreneurs building an AI business today

Web: https://www.fiddler.ai

Linkedin: https://www.linkedin.com/company/fiddler-ai/

Twitter: https://twitter.com/krishnagade

Join Fiddler: https://www.fiddler.ai/careers

At Dawn, we invest in visionary leaders who move the world forward. Learn more at https://dawn-gv.com    

With questions about Forward contact us at [email protected]

We caught up with Ankur to talk about all things ChatGPT.  

Ankur’s note: https://www.ankursnewsletter.com/p/nine-chatgpt-products-we-can-expect 

0:00 CGPT advantages over the previous generations of LLMs 

3:30 RLHF as the optimization method behind CGPT 

7:13 Products to expect 

12:55 CGPT challenges  

18:33 LLM landscape 

21:43 Will the number of parameters continue to grow 

24:25 The future of transformer architecture 

29:00 Opportunities for startups  

CNet discovers errors in more than half of its AI-written stories: 

https://www.theverge.com/2023/1/25/23571082/cnet-ai-written-stories-errors-corrections-red-ventures 

Mikela is the co-founder & CEO of Greyparrot, the leading AI waste analytics platform for the circular economy on a mission to increase transparency and automation in waste management and unlock the financial value of waste. Greyparrot’s vision is to create a world where every piece of waste is valued as a resource.  Greyparrot is recognized as one of “Europe's 100 Hottest Startups” by WIRED in 2021 and “100 most promising AI companies” globally by CB Insights.

Greyparrot is backed by Una Terra, Unreasonable Group, Speedinvest, Closed Loop Partners, TI Capital, and Plug and Play, among others.

0:00 The state of the waste management industry

6:14 Waste contamination, waste purity

9:01 Major challenges of the recycling industry

12:45 Greyparrot’s material monitoring system

18:00 How much is our waste worth?

19:34 Waste management market structure in developed economies

21:50 How to measure the success of Greyparrot’s implementation

25:14 Computer vision systems powering Greyparrot

34:34 Why apply computer vision to waste management

37:04 Finding product/market fit and recruiting design clients

45:32 Raising capital - creating a system around fundraising

54:25 Greyparrot’s culture: culture fit vs culture add

01:01:47 The future of the circular economy

Links:

Ellen MacArthur Foundation: https://ellenmacarthurfoundation.org/

Web: https://greyparrot.ai

Linkedin: https://www.linkedin.com/company/greyparrotai/

Newsletter: https://www.greyparrot.ai/stories

Work at Greyparrot: https://www.greyparrot.ai/careers

At Dawn, we invest in visionary leaders who move the world forward. Learn more at https://dawn-gv.com

With questions about Forward contact us at [email protected]

Our guest today is Daniela Braga, the CEO and founder of Defined AI, the world’s leading online marketplace, where AI professionals can commission, buy, or sell AI training data, tools, or models.

We sat down with Daniela to talk about building the largest data marketplace in the world, GDPR, China/US vs. the rest of the world, raising $80M+, partnering with IBM, working with IBM, Microsoft, and a “Cupertino-based company”, Working on the US Govt Taskforce to democratize AI, and building an AI hub in Portugal.

0:00 - Defined AI - the largest marketplace of data

09:25 - Is there a competitive advantage in data

12:20 - Leveraging marketplace strategy to double customer count, 6x revenue in a year

14:05 - Simulated vs synthetic vs live data

18:40 - How to crowdsource high-quality data

23:52 - The future of simulated data

28:03 - Challenges of using real data

32:35 - Powering IBM Watson, working with Microsoft and a “company based in Cupertino”

36:00 - From majoring in literature and linguistics to building AI systems and entrepreneurship

47:00 - Living in the US/Portugal/China

54:00 - What language is closest to math

55:30 - Working with the US Govt Taskforce to democratize AI

1:01:00 - Building an AI hub in Portugal

Links:

NAIRR: https://www.whitehouse.gov/ostp/news-updates/2021/06/10/the-biden-administration-launches-the-national-artificial-intelligence-research-resource-task-force/

Portugal AI Hub - Accelerat.ai: https://portugal.postsen.com/news/159065/Daniela-Braga-will-have-345-million-euros-to-put-Portuguese-AI-on-the-world-map.html

Main website: https://www.defined.ai/

Neevo platform: https://www.neevo.ai/

Daniela on Linkedin: https://www.linkedin.com/in/danielabragaspeech/

Twitter: https://twitter.com/dbraga51?lang=en

Careers: https://www.defined.ai/careers/

We sat down with Glean AI’s Ankur Patel to discuss managing business costs with AI, running an AI-powered hedge fund, culture at Bridgewater, building the remote-first company, and the future generative AI. 

Ankur is a co-founder and head of data at Glean AI. Glean uses natural language processing to deliver vendor spend intelligence within an accounts payable solution. Before founding Glean, Ankur was the vice president of data science at 7Park Data, a Vista Equity Partners portfolio company. 

Ankur used alternative data to build data products for hedge funds and developed a natural language processing-based entity recognition, resolution, and linking platform for enterprise clients. Before 7Park Data, Ankur led data science efforts in New York City for Israeli artificial intelligence firm ThetaRay, a pioneer in applied unsupervised learning. 

Ankur began his career as an analyst at J.P. Morgan and then became the lead emerging markets sovereign credit trader for Bridgewater Associates, the world's largest global macro hedge fund. He later founded and managed R-Squared Macro, a machine learning-based hedge fund. 

Ankur is the author of Hands-on Unsupervised Learning Using Python and Applied Natural Language Processing in the Enterprise, both O’Reilly Media publications. 

Glean is backed by Contour Venture Partners, Infinity Ventures, B Capital, Portage Ventures, Amex Ventures, and other prominent venture funds and fintech angels. 

The audio version is available everywhere you listen to podcasts: https://anchor.fm/forward-by-dawn/ 

0:00 - Why start a career on Wall Street 

03:55 - Key takeaways from Bridgewater's approach to culture 

05:55 - How to build an AI-powered hedge fund 

08:25 - Can a data edge be sustainable for a hedge fund 

10:35 - Data-driven hedge fund horizon 

13:30 - How should an individual investor think about the market and edge 

17:10 - How Glean helps organizations manage expenses with peer intelligence 

24:10 - Scaling spend analytics  

29:00 - Offering Glean functionality as a set of APIs 

36:00 - AI technologies powering Glean 

40:35 - Solving the “cold start” problem with training data 

43:30 - Raising a multi-million seed round pre-product and pre-revenue 

48:00 - Managing the growth, fighting the pressure of growth-at-all-costs 

52:20 - To build a strong culture focus on what motivates people 

56:00 - Adapting the culture to the remote-first environment  

1:00:00 - Think about how acquiring data makes a product better 

1:05:00 - The future of generative AI and multi-modal models 

Company: https://www.glean.ai/  

Twitter: https://twitter.com/aapatel09  

Newsletter: https://www.ankursnewsletter.com/  

At Dawn, we invest in visionary leaders who move the world forward. Learn more at https://dawn-gv.com   

With questions about Forward contact us at [email protected]

We sat down with Aaron to talk about building bots to compete in games, Raising a $4M seed in a week from NEA and A16z, building an AI-powered company, and the future of AI in gaming.  

Aaron is the Founder and CEO of Regression Games, a company making artificial intelligence in games and esports accessible and enjoyable for everyone. He was previously an early engineer at Instabase, working on low-code machine-learning products. He received his Bachelors and Masters from MIT, where he helped develop Battlecode, MIT's annual AI gaming competition.  

Start competing in Minecraft - Ultimate Collector challenge at https://play.regression.gg  

Timestamps:  

0:00 - Why let people build AIs that will compete in games 

2:10 - Which games to use as arenas for AI bots' competition 

8:10 - First tournament - Minecraft Ultimate Collector 

10:46 - RG approach vs. other companies applying AI in games 

16:40 - The toolset to program bots 

18:33 - Leveling the field for players with various programming skills 

21:00 - Origins: First robotics, MIT Battlecode, Instabase 

31:00 - Applying Natural Language and LLMs to programming bots 

35:10 - How to monetize AI bots 

38:15 - Starting RG as a solo founder, first steps, NEA, A16z, Roosh 

41:00 - Navigating the venture ecosystem as a first-time founder 

45:05 - Building initial team 

48:37 - Future of AI gaming 

52:25 - Advice to founders starting an AI company 

57:05 - Future of AI in general, AI as a companion   

Links:  

Regression Games: https://www.regression.gg  

Discord: https://discord.com/invite/925SYVse2H  

Twitter: https://twitter.com/RegressionGG   

At Dawn, we invest in visionary leaders who move the world forward. Learn more at https://dawn-gv.com    

With questions about Forward contact us at [email protected]

Our guest today is Saman Farid, the CEO and founder of Formic. Formic is accelerating the adoption of automation across small to mid-sized U.S. manufacturers and boosting productivity by removing the risk, complexity, upfront capital, and expertise that typically exist as barriers to onboarding this technology.

The company’s “Robotics-as-a-Service” model delivers customized robots at a low hourly rate, with guaranteed uptime. With a full-service offering that includes everything from planning and deployment to maintenance and upgrades, Formic is creating an innovative and accessible competitive advantage for small and mid-sized factories to ensure they stay relevant and thrive in any economy.

0:00 - Manufacturing in the US

4:15 - Why labor shortages

7:52 - Robotization in the US, China, and Korea

13:47 - Opportunity for the US manufacturing automation

19:08 - Formic’s hardware, software, and financing services

24:44 - Types of customers served by Formic

30:55 - Collaborative robots

33:15 - Case study: auto parts manufacturer

39:15 - Learnings from 10 years as a VC

46:56 - How to incubate a business

54:44 - People at Formic

1:02:12 - Future of robotization in manufacturing

TRIZ: https://en.wikipedia.org/wiki/TRIZ

Company: https://formic.co/

Twitter: https://twitter.com/goformic

Open positions: https://formic.co/careers

At Dawn, we invest in visionary leaders who move the world forward. Learn more at https://dawn-gv.com

With questions about Forward contact us at [email protected]

Our guest today is Tomas Vykruta, the CEO and founder of EvolutionIQ. EvolutionIQ created the world's first fully AI-driven claims guidance platform for the most sophisticated insurance carriers in the world, making insurance more fair, transparent, and affordable.

They raised more than $26M from Brewer Lane Ventures, FirstRound Capital, FirstMark Capital, and Foundation Capital, among others.

0:00 How does insurance claims guidance help insurers 

3:00 Insurance companies served by EvolutionIQ 

8:26 Customers' success stories  

10:45 AI systems powering EvolutionIQ 

15:37 Future roadmap 

17:50 Convincing your first customer to share the data 

21:26 How can an AI team become experts in insurance 

29:22 Starting EvolutionIQ 

37:57 Lessons from Google 

44:06 Lessons from building EvolutionIQ 

55:11 The future of AI in insurance

A Recipe for Training Neural Networks: http://karpathy.github.io/2019/04/25/recipe/

EvolutionIQ: https://evolutioniq.com/

Careers: https://evolutioniq.com/careers/

Linkedin: https://www.linkedin.com/company/evolution-iq/

Tomas’ Twitter: @tvykruta

Tomas’ Linkedin: https://www.linkedin.com/in/tvykruta/

At Dawn, we invest in visionary leaders who move the world forward. Learn more at https://dawn-gv.com

Our guest today is Mark Angel, the CEO of Amira Learning.  Amira is the first intelligent reading assistant that listens, assesses & coaches to accelerate reading mastery. Proven to double the pace of learning, Amira helps K-3 students become motivated, masterful readers.

The company recently debuted StoryCraft - a new reading game app for kids ages 5 to 10, which uses artificial intelligence to deliver personalized reading tutoring from the comfort of home.

Students have read more than 1 billion words to the software, with recent usage hitting 2+ million minutes per month, spanning nearly a million students in all 50 states.

Website: https://www.amiralearning.com/

Linkedin: https://www.linkedin.com/company/amira-learning

Careers: https://www.amiralearning.com/jobs

At Dawn, we invest in visionary leaders who move the world forward. Learn more at https://dawn-gv.com

With questions about Forward contact us at [email protected]

Our guest today is Ittai Dayan the Founder and CEO of Rhino Health.  Rhino Health provides AI/data-science developers and medical researchers with a platform that unlocks data collaborations without sharing data or compromising data privacy. The Rhino Health Platform is powered by Federated Learning and distributed computing technologies and uniquely adapted to deal with 'Real World Data' from clinical data systems.

In this episode, we talk about the challenges AI innovators face with gaining access to healthcare data, the tradeoffs of different approaches, and the opportunities of federated learning. We also talked about harmonizing the data from different institutions and avoiding the “data black box” while ensuring data protection.

Website: https://www.rhinohealth.com/

Linkedin: https://www.linkedin.com/company/rhinohealth/

Contact us at [email protected]