Tobias Zwingmann

RAPYD.AI

Independent AI consulting to create profitable AI solutions to innovate and grow your business. Fast.

Managing Partner

O'Reilly Media

Author and Instructor

LinkedIn

LinkedIn Learning Instructor

D2A2

Chief of Research - Applied AI

Springboard

- Helping Springboard as a subject matter expert in AI, especially for data analytics
- Mentoring students who want to switch careers to data science / data analytics 
- Top Mentor Award in 2021 (Springboard Inspire)

AI SME & Data Science Mentor

Insurgence

Partner

Toptal

Independent Consultant

Data Science Dojo

ML/AI Instructor

Deutsche Messe AG

Systematic retrieval, analysis and utilization of (unstructured) data from diverse sources to build new products and processes.

- People movement analytics 
- Computer vision with focus on text retrieval

Senior Data Scientist

Building & maintaining predictive models. Generating and communicating insights about customer trends and behavior to product owners and senior management.

- Churn Prediction
- Next Best Offer Recommendations 
- Customer Segmentation
- Visual Analytics

Data Scientist

Leading various international trade show projects in e.g. China, India, Indonesia, Mexico, Myanmar, South Africa, South Korea, Turkey.

Project Director

Deutsche Messe Dubai Branch

Management Trainee

Author of AI-Powered BI (O'REILLY) | AI Coach & Mentor | I help organizations leverage AI and use data more intelligently to improve business performance. Sharing AI use cases & learnings.

I didn’t start my AI/ML journey doing “Profitable AI”.

I started it as a Data Scientist who built models that looked impressive on paper:

– Customer churn prediction with 95% accuracy ✅
– Next-best-offer systems with 15% uplift ✅
– Data-driven B2B products generating 6-figure revenues from day one ✅

But the hard lesson I’ve learned after 10+ years at the intersection of business and technology is that none of this matters if you can’t show results on the bottom line.

I’ve spent the last 5+ years in applied AI consulting where my work has taken me inside global enterprises in automotive and insurance, mid-sized companies with 500–1,000 employees, solopreneurs aiming to double their output, and small business owners uncovering $10k+ monthly savings in unexpected places. 

The contexts differ, but the patterns of what makes AI succeed (and fail) remain remarkably consistent.

Success is not about:
- Chasing the next shiny AI tool
- Hiring data scientists from Silicon Valley
- Hoping IT will figure it out

It’s about applying the principles of Profitable AI to deliver outcomes that matter: doing things faster, at lower cost, with higher quality — or scaling what you do to entirely new levels.

Along the way, I’ve shared many of these lessons in my books AI-Powered Business Intelligence (O’Reilly, 2022) and Augmented Analytics (O’Reilly, 2024), as well as in my weekly newsletter that is now read by 10,000+ professionals at companies like Google, Amazon, Gucci, H&M, Santander, and Mercedes-Benz.

So If you’re tired of the shiny AI demos and want to see it deliver real results instead — I think you found a good place.

We help you build amazing AI products faster.

Gemini 3 is making a lot of noise. And it's a really great model. "Is it better than GPT-5?!" My take: Use whatever works for you. Both are LLMs, both have their quirks. But if you want to know where each shines, here's what I found:

Gemini 3 = Messy Inputs → Clean Outputs
Throw it chaos: screenshots, transcripts, logs, mixed documents. It's great at compressing that mess into structure. Best when you're very specific about the output format you want. Seems like if you need to organize messy information, call Google.

Examples:
- Summarizing multiple meeting recordings into structured notes
- Extracting data from a messy pile of PDFs, screenshots, and spreadsheets

GPT-5.1 = Clean Inputs → Complex Thinking
Give it organized information and let it reason. It's great at planning, narrative, and multi-step decisions. Best when the task itself is complex, not the input.

Examples:
- Writing code (though Claude still beats it here)
- Building a strategy doc from a clean brief

Scenarios to combine both:

1/ Competitive analysis
Gemini 3 gathers data from competitor sites, pricing, and reviews; GPT 5.1 delivers the strategy

2/ Client onboarding
Gemini 3 cleans and organizes intake docs; GPT 5.1 produces the project plan

3/ Sales pipeline review
Gemini 3 compiles CRM data, calls, and emails; GPT 5.1 spots trends and recommends actions.

Don't try to find "the" best. 
Find whatever works best for you.

Fun fact: The infographic was done by me with Gemini-3 image gen ("Nano Banana"), first shot, Prompt in comments.

AI Advisor | O’Reilly Author | Instructor @ LinkedIn Learning | I find & realize profitable AI opportunities – here to share wins, fails, examples & frameworks.

I see a lot of confusion around AI workflows, AI agents, workflow automation, and agentic workflows. Most people can't even tell the difference. Which isn't a problem until you want to build an AI solution that is neither over-engineered nor insufficient.

So I put together a quick guide that breaks down each type and when to use what, along with a few examples.

Read the full blog post here: https://lnkd.in/erYqeVPi

AI Workflows vs. AI Agents

ChatGPT launched today 3 years ago. ONLY three years I should say. Feels like a decade compressed. To catch you up on what’s happened, I generated this little timeline with the help of ChatGPT Deep Research, had Gemini 3 double-check it, and Claude 4.5 generated the PDF.

Tasks like this would have taken a whole day in the "old" world.
Now done in less than 60 minutes with AI.

We're still far away from AGI.
But I don't care.
We're living in the most exciting timeline.
And I can’t wait to see what this timeline looks like 3 years from now.

3 Years Since ChatGPT

A while ago, I shared the AI Profit Stack – a simple framework to assess whether your AI projects are actually creating profit. People loved it, clients used it. Then someone asked me:

“How do I apply this to my Copilot licenses?”

Short answer:
You don’t.

Long answer:
Trying to measure all AI with the same ROI formula is exactly how you end up with fictional numbers, cynical finance teams, and AI steering committees that debate themselves into extinction.

Today I've shared the fix I'm using now in every client engagement – and the one I want you to have too.

Find it here here:

Open article: The Two Tracks of AI ROI by blog.tobiaszwingmann.com

How to separate 'AI that helps' from 'AI that pays'

🎯 Mark Freeman II

Gable

• Joined as employee-one to lead GTM at Gable to help us reach our first customers.

• Reported directly to the CEO and led the development and execution the go-to-market
strategy which resulted in a successful launch, 4000+ leads, and 300+ meetings with
senior data engineers and leaders, ranging from scaleups to Fortune 500 companies—
this was integral for our ability to raise a Series A. (https://bit.ly/4eDwbPE)

• Co-authored the O’Reilly book "Data Contracts: Developing Production Grade Pipelines at Scale,"
contributing all the technical chapters, and positioning Gable as thought leaders of an
emerging category within the data market. (https://bit.ly/3XIu4mF)

• Served as a technical expert by presenting at conferences on data engineering topics and providing thought leadership via blogs.

• Built Gable’s marketing and sales stack with an emphasis on data capture, automation,
and analytics that enabled automated qualification, conversion into meetings, and nurture
campaigns for leads—while establishing the underlying data model.

• Served as a sales engineer, mapping technical requirements to Gable’s product roadmap,
building champions among directors plus data leaders, creating data strategy decks for
clients, and closing multiple accounts.

• Served as a data scientist, providing insights and visualizations for Gable’s executive team
for presenting sales metrics to the board and potential investors.

Tech Lead, GTM Engineering

On the Mark Data

I help brands connect to data professionals via content strategy, GTM strategy, and influencer marketing. Learn more at onthemarkdata.com.

• Built a brand with an audience of 40k+ data professionals via social media content, technical blogs, and video-- which has harnessed over 14MM impressions.

• Established the weekly Scaling DataOps newsletter, with 7k+ subscribers, where I interview data leaders and executives on data best practices in scaling data infrastructure.

• Built a revenue stream around sponsored content from top data brands such as Nvidia, SuperDataScience, Scale AI, and much more.

Founder

O'Reilly

Currently writing a book on data contracts.

⭐️ "This book has the potential to become the definitive guide for data contracts! I can't wait to dive into the rest of it and unlock even more valuable insights!"

⭐️ "The initial read is clear and provide a good potted history of the data problem and how it has organically morphed due to so many factors. I can't wait for the rest of the book"

⭐️ "This book has the potential to become the bible of data contracts especially as the world moves towards intelligent applications. Data is essential to applications that use AI, and having a good contract is paramount towards establishing trust between systems and having a seamless experience. Looking forward to the full publication!"

Author - Data Contracts: Developing Production Grade Pipelines at Scale

⭐️ "Took a few hours out of my Saturday to dive into data engineering with DBT, and had the opportunity to pair program with Mark... Great work on the new course, Mark. I sure learned a lot, and I'm confident others will benefit from it as well. As someone who's relatively new to data engineering, this short course was an eye-opener. It expanded my understanding of the scalability and possibilities within the field."

⭐️ "I completed Mark's course over the weekend and thought it was a GREAT introduction to using DBT. For people looking for a low-stakes, easy-to-follow introduction to DBT as a tool for data engineering, I think you would be pressed to find something more effective."

⭐️ "I would like to thank Mark... for creating such a comprehensive, engaging & pure hands on course on dbt. He explained the concepts clearly & demonstrated them with practical examples and exercises."

⭐️ "I was excited when I saw this course on your page, i know this is going to be the course I needed to start my DBT journey and you did not disappoint, Thank you"

LinkedIn Learning Instructor, Data Engineering

Humu

• Build ETL pipelines to bring third-party data sourced from APIs to combine with our product data, creating rich data assets to drive product feature creation and customer analytics.

• Built buy-in, developed the transition plan, and led the migration of the orchestration of data transformations from manually created scheduled queries to data build tool (dbt)-- which enabled version control of our semantic layer, data quality checks, and data discoverability of data assets within our data warehouse.

• Led the development of an internal data quality engine to increase test coverage of core data pipelines and implement data observability.

Senior Data Engineer

Driving forward the organization’s data strategy by:
• 1) Conducting user interviews to understand the organization’s data pain points.
• 2) Identifying the need for data marts to increase data access.
• 3) Scoping overhaul of our data pipelines and warehouse.
• 4) Sourcing data ingestion vendors.
• 5) Fostering leadership buy-in via building minimum viable products.

Senior Data Scientist (Data Engineering)

• Led the development of the Data Science team’s custom R package and managed the internal distribution and upgrading.

• Implemented CI/CD utilizing Docker and CircleCI to ensure changes to the Data Science codebase passes tests and prevent breaks to our internal packages.

• Assisted in doubling the Data Science team via interviewing, onboarding, and technical mentorship for upskilling colleagues.

• Developed and put into production an NLP pipeline via Python that takes raw text data, classifies respective texts' themes, and ranks and assigns text to relevant insights within customer reports (https://bit.ly/3yAeBqu).

• Designed and put into production via Python an algorithm to classify user activity—output became org’s main product health metric and was validated with both statistical significance and UX research interviews (https://bit.ly/3yt5QhY).

• Support the VP of Product in strategic product decisions by refining business questions, conducting log analytics via SQL and R, and creating reports shared org-wide.

IndexoRx

• Made it to the final round of interviews for Y Combinator, where we presented our pitch deck to their investment team and described the business opportunity within the pharmacy space in reducing the burden of DIR fees.

• Developed an MVP of our data product that utilized analytics and machine learning to rank users at risk for lowered medication adherence and forecast DIR fees.

• Conducted deep market research on pharmacy business models, pharmacy supply chains, and PBM reimbursement processes.

Founder, Chief Data Officer

Bay Area Community Health Advisory Council

• Lead organization’s evaluation efforts through developing a standardized survey utilizing CSAT and NPS scores, transitioning administration from paper to Google Forms, and providing analysis and data visualizations of services data—resulting in a 24% increase in survey response rate and 2 reports used by org’s Executive Director for fundraising

Member / Evaluation & Analytics Committee Chair

Verana Health

• Utilize deep understanding of medical claims and EHR data, in conjunction with strong SQL and PySpark skills, to work with data that contains over 63 million patients, 274 million patient encounters, and over 10 billion data points amongst 55 separate EHRs

• Developed and deployed into production an algorithm via Apache Spark that classified insurance types for over 250 million records, utilizing unstructured EHR data, that improved both accuracy from 40% to 84% and reduced null values from 16% to <1%

• Support sales team in 21 potential contracts by determining viability of our data assets through defining research and business questions, developing SQL and PySpark code to build cohorts, providing insights on quality issues within data, and creating data reports to communicate findings to non-technical staff

• Conduct an algorithm assessment on a key variable and identified errors impacting 1.3 billion records, as well as provide recommendations to the algorithms team to rectify the error—assessment was utilized to help certify new data architecture

Data Scientist (Data Quality)

Stanford University School of Medicine

• Developed R script for department’s ETL process, that transforms data into cleaned variables and removed duplicate observations based on defined logic—reducing data cleaning time of analysts by 10 hours per project

• Completed an exploratory analysis through conducting 48 multiple regression models to understand relationship between biometric data and self-reported wellbeing data, thus identifying 3 viable hypotheses to further investigate and ensuring the lab diverts resources to impactful studies 

• Updated methods of a previous propensity score match analysis, lowering the SDM for all variables from 0.17-0.21 to <0.10, resulting in the analysis now meeting required match assumptions and improving the manuscript strength for publishing: matching accounts for observed confounder bias

Data Analyst

Collective Health

• Developed and operationalized a Python script, utilizing pandas, that identifies discrepancies between insurance enrollment data and network partner data— reduced manual work process time from 98 to 16.3 hours a year (time saved grows with number of added clients)

• Developed and operationalized a Python script, utilizing pandas, that takes 3 sources of data and creates custom reports for respective insurance partners’ specification— reduced manual work process time from 52 to 4.3 hours a year

• Utilize SQL to pull necessary data to resolve insurance eligibility issues and other related operations work

Operations Associate

onthemarkdata.com | Senior data engineer helping brands with GTM and content strategy.

I'm a community health advocate turned data engineer interested in the intersection of social impact, business, and technology. My life’s mission is to improve the well-being of as many people as possible through data—especially among those marginalized.
 
As the founder of On the Mark Data, I use this platform to help brands connect to data professionals through captivating content.
 
I received my M.S. from the Stanford School of Medicine and I'm also certified in Entrepreneurship and Innovation from the Stanford Graduate School of Business. In addition, I have worked within numerous startups where I have put machine learning models into production, integrated data analytics into products, and led migrations to improve data infrastructure.
 
The culmination of my technical skills, content success, and experience within startups would make me an asset whether you are looking to create a successful content strategy, engage in influencer marketing campaigns, or develop a go-to-market strategy for your startup.

You can learn more about opportunities to collaborate, my media appearances, or my content at onthemarkdata.com.

• Joined as employee one to lead GTM at Gable to help us reach our first customers.
• Worked with the CEO to establish our GTM strategy, content strategy, and messaging for technical audiences.
• Building internal data products to drive our GTM strategy and sales.

It didn't feel real until I had the physical copy of the book in my hands, but I'm officially a published author!

This is just such a full circle moment for me as I was reading these O'Reilly animal books when I was dreaming of making a career in data.

If you told me nearly 7 years ago that I would have my name on one I wouldn't believe you!

This represents about two years of life of nights and weekends on top of helping build a startup (Gable) and my own company (On the Mark Data).

Writing a book and building a company was hard... but I would easily do it all over again.

Hundreds of user interviews and discovery calls informed us what would be meaningful in the book.

While the book provided us clear synthesis of what we were seeing in industry.

This created a symbiotic relationship where we would build, write, get feedback, and repeat.

A huge shoutout to the O'Reilly team and my co-authors Chad and Ben!

An even bigger shoutout to everyone who supported us, as it wouldn't have been possible without you all.

Data Engineer | Tech Lead @ Gable.ai | O’Reilly Author: Data Contracts | LinkedIn [in]structor (30k+ Learners) | Founder @ On the Mark Data

No alternative text description for this image

Here's my tech stack for the dev environments I create for LinkedIn Learning courses (all open source) that you can use for yourself:

🐋 Dev Environment:
- Docker Compose - Containers and configs
- VSCode Dev Containers - IDE in container
- GitHub CodeSpaces - Browser cloud compute

📦 Databases:
- Postgres - Transactional Database
- Minio - Data Lake
- DuckDB - Analytical Database

⚙️ Ingestion + Orchestration + Logs:
- Python Scripts - Simplicity over a tool
- Data Build Tool - SQL queries on DuckDB
- Alembic - Python-based database migrations
- Psycopg - Interact with postgres via Python

📋 CI/CD:
GitHub Actions - Simple for students
 
📊 Data:
Data[.]gov - Public real-world datasets

👩🏽‍💻 Coding Surface:
Jupyter Notebooks - Quick and iterative
VS Code - Update and implement scripts

This setup is extremely powerful as you have a full data platform that sets up in minutes, it's filled with real-world data, you can query it right away, and you can see the logs. Plus, since we are using GitHub codespaces, it's essentially free to run in the browser with just a couple clicks!

Innovation isn't about creating completely brand new technologies and practices. Instead, it can be applying a proven method to a new domain.

Under this framing, data engineering teams can learn a lot from clinical research methodologies.

My stats professor in grad school used to joke around that someone messed up on research design if he had to use advanced statistics.

Why? You can avoid jumping through so many technical hoops if you control for people and process!

Here's how I see it applying to data systems:

A - clinical) How do we ensure fidelity (ie research staff follow protocols correctly) across the steps of data capture for the study?

A - data) How do we ensure the business processes for capturing and processing data stays consistent across staff (eg data entry)?

B - clinical) How do we store patient data, and what is the least amount of information we need to reduce risk for a HIPAA violation?

B - data) What data values are considered PII, how do we retain and delete such records, and what regulations do we need to handle?

C - clinical) What are we trying to compare (eg treatment vs control), and thus what measurements and demographics do we need to capture to properly assess this?

C - data) What telemetry do we need on a product to answer questions about usage of a product or service?

Nine times out of ten, the problem you are solving for already has a solution in a different context!

Stop looking at problems at a surface level and instead understand the "pattern" of the problem, and the find similar patterns in spaces where the problem has both been solved and well documented. It's why interdisciplinary teams are the most innovative!

This is also exactly how we landed on "shift left data" in our data contract book by the way! We saw the same problem pattern in DevOps and how they used "shift left" to address it.

"What do you mean my ROI would only be 1.5x? I need it to be a 4x for it to make sense..."

That brutal honesty was one of the best lessons I've learned in getting adoption and budget for data infrastructure projects.

I spent days making a proposal for a c-suite exec I was trying to sell to. I identified their core problems, was getting multiple meetings, and we were talking about numbers... and that's where I messed up.

My mistake was assuming this was their only problem they were considering for purchasing, and not part of a portfolio of problems they were strategically trying to solve.

I wasn't "competing" against a problem I knew I could clearly solve and generate ROI on. Instead, I was competing with the opportunity cost of not pursuing otherr solutions with various multiples of ROI for different problems.

That meeting completely changed how I think about value to execs. It's funny how obvious it sounds looking back, but I'm hard headed and needed to learn by experience.

Fast forward to today, and I feel like I have had enough reps to dive into the business technicals of what value actually means and map it to my experiences.

This week I'm reading the book Valuation, which was essentially the playbook used by McKinsey consultants. Specifically, I want to take my understanding of what is valuable to a business to next level with economic theory and technical business fundamentals.

I think more technical people should study this as well! Move away from parroting the abstract phrase "we need to focus on driving business value" and instead provide a clear plan to what that actually means and why a leader shoukd choose to invest in your initiative over others.

Allowing others to explore AI >>> Gatekeeping AI

I still remember the first time I saw a neural network work in real-time (shoutout the MarI/O Youtube video by SethBling) and just the magic I felt.

That was the moment I realized I needed to make a career in data and the moment I got over my fear of coding.

So many non-technical, but highly skilled in their domain, professionals are now experiencing exactly what I felt years ago.

Why try and police that?

Yeah, that vibe coded app is absolute crap and one agentic action away from deleting its own databases... but that person shipped something.

That's already more progress than the technical naysayers shouting from the sideline.

taplio

The Battle

Tobias Zwingmann

🎯 Mark Freeman II

Followers

Posts

Engagements

Likes

Performance ScoreTobias Zwingmann's LinkedIn Performance

Performance Score🎯 Mark Freeman II's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Tobias Zwingmann

🎯 Mark Freeman II

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Tobias Zwingmann's LinkedIn Performance

Performance Score🎯 Mark Freeman II's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreTobias Zwingmann's LinkedIn Performance