Adam Sroka

Data & AI Director

Hypercube Consulting

Chief Executive Officer

Director

The Data Lab - Innovation Centre

The Innovation Advisory Board is responsible for advising on innovation and collaborative project activity for The Data Lab – Scotland’s Innovation Centre for Data and AI.

Innovation Advisory Board Member

Origami

Head of Machine Learning Engineering

Incremental Group

Senior Data Scientist

Sure Thing! Insurance

Data Scientist

Clear Returns

Thales

Research Engineer

I build data and AI platforms for the energy sector | Director at Hypercube Consulting

Dr Adam Sroka is a Co-founder and CEO at Hypercube Consulting, a firm that helps organisations unlock value from data by delivering enterprise-scale solutions and building high-performing data and analytics teams from the ground up. With over 10 years of experience in data science, machine learning, and artificial intelligence, Adam is a recognised expert and speaker in his field.

Adam's mission is to bridge the gap between technology and business, and to enable data-driven innovation and transformation across various industries. He shares his insights and best practices through his blog, podcast, and tech community events, where he showcases his approaches and systems for effective communication, engineering discipline, and experimentation. Adam is passionate about helping organisations and data professionals maximise their potential and impact with data and AI.

botnanza

> Helping robotics companies to build digital presence & global brand awareness in media platforms.

Chief Evangelist

ff Venture Capital

> Identifying seed / series A robotics companies.

Venture Partner

Various Companies

> Investing in early-stage startups in robotics market

Angel Investor

First Momentum Ventures

> Identifying first-check, pre-seed deep-tech companies (industrial software / robotics)

Warsaw University of Technology Business School

Visiting Lecturer

Energy Robotics

> Creating an end-to-end AI-powered platform for inspection robots & drones.

Global Business Development

KOGENA - Robots Monitoring Platform

> Diagnostics and predictive-maintenance platform for cobots and industrial robots.

Business Development Executive

ProCobot

> Driving sales & business development in southern Poland - cobots (Universal Robots), mobile robots (MiR) & service robots (Pudu Robotics).

> Introducing service robotics (restaurants, hotels) to the Polish HoReCa market.

Sales & Business Development

Software Engineer

WAGO

> Designing control systems including telemetry, PLCs & IoT devices.

IoT & Automation Trainee

Robotics evangelist @ planet Earth 🌍

My LinkedIn brand is #1 worldwide in the robotics space. I work with some of the largest robotics companies in the world. I travel a lot (I love my job 💛)!

A content creator who is obsessed with technology, especially robotics & AI/ML! 🚀 

Basically, I help high-tech companies with digital enablement (creating an audience in the digital world, generating leads, and building brand awareness).

Who am I? 

→ 1st World Content creator on Linkedin (+180k subscribers) in Robotics 🤖 & AI 🧬
→ Elon Musk reposting an commenting my robotics news on X (+35M views annually on X)
→ University Lecturer in the field of Industry 4.0 technologies 🎓
→ +85M views annually on LinkedIn 📈

Twitter: 
https://twitter.com/lukas_m_ziegler

Be part of my community for free at → ziegler.substack.com

📬 Send a DM for business opportunities.

Lukas M. Ziegler

Robot dogs join firefighting crews! 🧯

In Sichuan, China, a quadruped robot just joined the firefighting team.

Deployed on the front lines, the robot assisted crews by scouting hot zones, carrying hoses, and operating in hazardous areas that would normally put responders at risk.

The quadruped platform combines mobility, thermal imaging, and remote operation to navigate unstable terrain and deliver real-time situational awareness.

Firefighting may be one of the toughest environments for automation, but every safe mission proves the same point: robots belong where humans shouldn’t have to go. 

~~

♻️ Repost to help 1 robot find a new workplace!

Robotics evangelist @ planet Earth 🌍 | Telling your robot stories.

A new data engine for Chinese humanoids! 🥢

Unitree Robotics has released their new system, a full-body teleoperation platform where a human in a motion-capture suit controls a G1 humanoid robot in real time.

Do you remember 'Real steel' with Hugh Jackman? 🎬

Yeah? Cool, so you know how it works.

The robot mirrors every move with near-zero latency playing football, sparring, and even boxing against another humanoid. 

But behind the spectacle is something deeper: a data-generation platform for physical AI.

I think that's the system’s real purpose: collecting high-quality motion data that maps human behavior directly onto robot kinematics.

Instead of relying on video or large-scale simulation, Unitree is betting on teleoperation as a data pipeline, using humans-in-the-loop to generate training data that’s already physically correct for the robot’s own body.

By combining an accessible hardware platform (the G1) with a sophisticated data capture approach, Unitree is on the way to own both software and hardware.

That's a huge advantage.

What's your opinion?

~~

♻️ Repost to help 1 robot find a new workplace!

Very sad news in robotics world. 😔

K-Scale Labs shuts down.

First of all, huge credit to them for their contribution to open-source and promoting DIY robotics.

Looks like they didn't secure the funding in time. Bit more context below.

Palo Alto-based humanoid robotics startup K-Scale Labs has officially shut down operations and refunded all preorders for its open-source humanoid robot, K-Bot.

Founded in 2024, K-Scale aimed to create a mass-produced, affordable, open-source humanoid, priced under $10,000. The compact 1.4-meter robot had attracted significant early interest, with 100 preorder slots and a few initial units shipped.

In an email to customers, their CEO Ben Bolte cited difficulties raising capital in the U.S. compared to well-funded Chinese competitors such as Unitree, Booster Robotics, or EngineAI, which have already moved into mass production.

Sad to see that, but I think they did an enormous work towards making humanoids and robotics in general more accessible! 🦾

~~

♻️ Repost to help 1 robot find a new workplace!

What's physical AI in reality? 🤨

When I visited Dexterity in Bay Area earlier this year, what struck me most was their approach to AI.

It isn’t one big, general-purpose system.

Instead, it’s a collection of specialized systems, each handling a different part of the overall task, loading and unloading trucks.

Dexterity, Inc.’s latest system uses multiple 3D cameras running multiple AI models at once, so robots keep working even when some cameras are blocked by equipment or movement.

Instead of one large neural network, the company uses a set of smaller, specialized models, for vision, motion planning, and force control, that work together in real time. 

The robots can sense force and pressure through their hands, letting them adjust their grip or movement as needed. Each arm and workspace is continuously mapped as a digital twin, helping the system avoid collisions and plan motion on the fly.

Packing tasks are optimized through real-time spatial analysis, improving trailer and container loading efficiency by up to 30% compared to standard pallet stacking.

It’s a focused approach to AI in robotics, not one model trying to do everything, but many systems working together to make robots more adaptable, precise, and safe in real operations.

It was one of the coolest robotics visits for me. Thanks for making this possible Samir Menon, and Michael Patrick Perry (I still remember that interesting story you told about how pallets were invented during our lunch!)

Super curious what are you going to show later today!

~~

♻️ Repost to help 1 robot find a new workplace!

I was crowded by robots! 😱

The Quality Group scales up with Locus Robotics bots.

At its new 40,000 m² facility in Elsdorf, the company is rolling out a fleet of Locus Robotics autonomous mobile robots, powered by the LocusOne platform. 

The goal is simple: scale up efficiently, handle rising order volumes, and move smarter.

The Locus system supports high-speed, high-volume picking while staying flexible enough to adapt to changing workflows. Robots can manage both fast-moving SKUs and slower, irregular items without rigid pallet setups or workflow redesigns.

Built for growth, the site is already equipped to handle up to 60,000 orders per day.

The bots run with FastPick, making them agile enough to manage peak periods, promotions, and influencer-driven demand spikes.

It’s a perfect example of the next phase in logistics automation! 

Congrats Sarah Hilgers and team for adopting technology that fast! 🔥

Good job Denis Niezgoda and crew for making this possible!

~~

♻️ Join the weekly robotics newsletter, and never miss any news → ziegler.substack.com

Well, it's Saturday.

A perfect moment to watch a robot giving you a new haircut, right? 💇🏼‍♂️

This robot created by Shane Wighton will not only do a great haircut or trim the sides but is also able to chat with you.

During the pandemic, the creator of StuffMadeHere built this robot to give himself a haircut. It offers different types of hairstyles and uses scissors. Parts of the robot were printed using a 3D printer. 

Fantastic design, but further down the line I'm wondering if I'm brave enough to let the robot give me a trim.

Probably not yet.

I would like to know more details about the safety and reliability of the system.

P.S. What are your honest thoughts on this project?

~~

♻️ Repost to help 1 robot find a new workplace!

🚨 BREAKING:

Another unicorn company in robotics industry! 🦄

At least from pre-liquidation event valuation persepctive.

Physical Intelligence just raised $600M at a $5.6B valuation!

One of the biggest rounds in robotics AI this year, and a clear signal of where the market thinks Physical AI is heading.

Physical Intelligence is building the software stack that lets robots learn complex, contact-rich tasks from demonstrations, corrections, and experience, the same approach behind π*0.6’s long-horizon coffee, laundry, and box-assembly performance.

On the video, you can enjoy not sped-up process of preparing an americano! ☕️

CapitalG led the investment, with Jeff Bezos jumping in alongside Thrive Capital, Lux Capital, Index Ventures, and T. Rowe Price.

Pretty neat amount of money to prepare for real-world deployment across physical domains (manufacturing, logistics or general industry). 💰

Is it the best week for robotics ever?!

~~

♻️ Join the weekly robotics newsletter, and never miss any news → ziegler.substack.com

A nano-factory in a box! 🥡

A system called MicroFactory has been shown autonomously assembling a photo frame, an actual product sold on Amazon.

Instead of relying on a massive general AI model, the team broke the job into 27 discrete actions: 23 recorded trajectories and 4 small, fine-tuned AI models for steps involving uncertainty.

When motion is predictable, like flipping a frame in a holder, the robot simply replays a recorded path. 

When variability appears, like peeling off a protective film, the AI model takes over.

It’s a simple but powerful approach: AI for randomness, recorded motion for repetition. Together, they handle the full assembly reliably and at low cost.

Hybrid methods like this can close today’s capability gap, collect high-quality data, and make automation economically viable.

~~

♻️ Repost to help 1 robot find a new workplace!

A new home-robot contender just dropped, and it’s all about the hands! 🤯

Tangible revealed Eggie, a wheeled humanoid that doesn’t settle for simple grippers.

 It’s running full five-fingered, anthropomorphic hands designed for real, contact-rich tasks.

In the demo, Eggie wipes a coffee spill in a real kitchen, not a staged lab scene,  showing off the company’s bet that true in-home robots need human-like dexterity, not shortcuts.

CEO Bipasha Sen is clear: the world is built for humans, so robots will need human hardware.

The philosophy? No real AGI without touch.

Tangible is gathering that touch data the extreme way, engineers wearing sensor-packed exoskeletons to capture real tactile interactions, from gripping a glass to pressing on a milk carton. 🥤

The home is the long-term goal, but the company is starting with the AI foundation, and offering $1M+ comp to recruit top researchers.

The wheeled-humanoid race was already heating up. Now there’s a new player betting that dexterity wins the home!

Thanks Hubert Thieblot & Founders, Inc. for backing such great projects! 🔥

~~

♻️ Join the weekly robotics newsletter, and never miss any news → ziegler.substack.com

Simulation as a data engine! 👾

Lightwheel is treating simulation as a data engine. 

The thing that's so special about them is the fact they can include physics and behaviours of objects.

The company’s pipeline measures real-world forces, mirrors them in sim, and scales learning through a human + AI + engineering loop.

Its stack is 90% automated, built on Blender human-in-the-loop QA. Every simulated contact is benchmarked against force/torque rigs and parameter sweeps, tuning friction, compliance, and material behavior.

The product line covers SIM-ready assets (rigid, articulated, deformable), synthetic data collection in VR/AR via teleoperation, and evaluation suites for closed-loop testing and reinforcement learning fine-tuning. 

A team of 300+ operators gathers real and egocentric human trajectories for robotic foundation models.

It's like a factory for data collection! 🤯

Demand is strongest from large AI labs, NVIDIA, OpenAI, ByteDance, Alibaba, looking to generate synthetic training data and RL fine-tuning runs across thousands of GPUs. 

Great to see more and more ambitious people tackle the data gap problem, let's go Steve Xie, Ph.D.! 🦾

Also a great robot demo experience at CoRL that included LeRobot teleops on deformables, all in simulation!!

~~

♻️ Join the weekly robotics newsletter, and never miss any news → ziegler.substack.com

Giving robots eyes on steroids! 👁️

Hello from Oslo, inside the headquarters of Zivid, where the team just launched their most powerful 3D camera yet: the Zivid 3.

During my chat with Henrik Schumann-Olsen, their co-founder CTO, we discussed one of the toughest challenges in logistics right now: depaletization of high-mix SKUs, those chaotic, real-world pallets full of differently shaped, reflective, or wrapped products. 📦

Zivid 3 was built exactly for that. It’s an 8-megapixel, large-volume 3D camera delivering millimeter-level accuracy up to 4.5 meters, with a capture time of just 500 ms. 

It can handle everything in a single scene — transparent film, shiny metal, dark packaging, and still produce consistent data across the entire pallet.

The challenge, as Henrik explained, is not just seeing the top of a two-meter pallet but maintaining the same image quality all the way to the bottom. 

That means managing light, glare, temperature shifts, and sunlight, conditions that typically break vision systems. ☀️

Zivid 3 solves this with robust lighting, temperature stability, and protection against ambient interference, giving integrators the performance they need for industrial depalletizing, bin picking, and any large working-volume application.

Or as they like to say at Zivid: “See everything. Do anything. From anywhere.”

~~

♻️ Repost to help 1 robot find a new workplace!

Back on stage! 🎙️

I haven’t performed in Poland for quite some time, it feels great to be back! 🇵🇱

It’s especially nice to be talking about robots here, in my 'own backyard'.

It's an event dedicated for system integrators, so the companies that are behind the creativity and outstanding engineering in all of these great robotic applications in industry.

These are the real builders! 🔥

In manufacturing, Poland had around 18,000 robots installed, which ranked it 6th in Europe and 17th globally, though again the density remains low.

Poland has tremendous potential when it comes to robotization. 

If you were to drive from Warsaw to Cracow, beyond the beautiful landscapes you’d see warehouses, production halls, and distribution centers, all places where robots can make a real difference.

That’s why conversations like this are so important.

I’m excited to be speaking at an event organized by KUKA, where I’ll discuss the future direction of robotics, including:

→ emerging trends in robotics innovation,
→ the importance of dexterous grasping,
→ how we approach autonomy, and locomotion
→ what are the emerging markets in robotics.

Thanks for inviting me Łukasz Szczepkowski, Artur Główka, let's rock! 🎸🧡

~~

♻️ Repost to help 1 robot find a new workplace!

That robot demo was so good, people thought it's fake! 🤯

XPENG had to prove its Robot wasn’t a human.

During their robot demo in China this week, the company’s humanoid robot Iron moved so naturally that people online started questioning whether it was real, or a person in a suit.

To settle it, engineers literally cut open the robot’s synthetic “skin” on stage, showing the mechanical structure underneath. 

Iron kept walking as if nothing happened.

Iron can perform 82 motions, including 22 degrees of movement per arm, with a design focused on realistic joint rotation and balance. 

XPENG says the robot will enter commercial use in 2026, aimed at shops and offices, not homes, where unpredictable environments still pose safety risks.

That's aura farming by the Chinese EV company! 🔥

~~

♻️ Join the weekly robotics newsletter, and never miss any news → ziegler.substack.com

taplio

The Battle

Adam Sroka

Lukas M. Ziegler

Followers

Posts

Engagements

Likes

Performance ScoreLukas M. Ziegler's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Adam Sroka

Lukas M. Ziegler

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);}Lukas M. Ziegler's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreLukas M. Ziegler's LinkedIn Performance