Saquib Alam

Building AI enabled hardware and software. founder at TheFluxTrain and General robotics @KTHuniversity grad, @iitbombay

[Saquib Alam @saquiboye]

It is surreal to see the product transform, from an basic AI model training platform to what it is today and crossing $5000 revenue this month, bootstrapped, run by a single person.

It took me 4 months to reach this milestone, since launch. But in reality, but the real journey began 24 months ago when I decided to return to India.

In that time, I’ve worked on multiple projects—some solo, some with a group. Some brought in zero revenue, while others made a few hundred dollars. It was a period filled with uncertainty, long months with no income, and much-needed breaks in Auroville to recharge.

Fast forward to today, TheFluxTrain is growing smoothly. The tech infrastructure is automated, lean, and designed to scale with just a small team/single person.

Unlike many other SaaS companies from India, a majority of revenue came from India(>50%), this month.

If my story can help or inspire others building in India (for the world), I’d love to share more insights. Let’s grow together. 🌱

And oh, Happy New year in advance.

Here is how my Stripe account looks like -

[Saquib Alam @saquiboye]

Umm, how is this even possible??

[Ronald van Loon @Ronald_vanLoon]

Breaking Gravity: Wall- and Ceiling-Climbing Warehouse #Automation
via @ZappyZappy7

#Robot #MachineLearning #ArtificialIntelligence #ML

[Saquib Alam @saquiboye]

Watching!

[Shubham @shubhdotai]

I'm spending the next 12 weeks building an Ascento-style robot from scratch - inspired from @mondorobotics

Target features:
self-balancing, fast acceleration, jumping, fall recovery, variable height, and all-terrain mobility.

I know nothing about Robotics, starting from scratch as an AI Engineer. I'll share everything I learn along the way.

#buildinpublic

[Saquib Alam @saquiboye]

Robots gaining skin in the game 🤪

[Chris Paxton @chris_j_paxton]

Giving robots "skin" which allows them to perceive the world by touch is important for improving task success, speed, and reliability in a wide range of different domains. DexSkin is an open source, conformable robot skin, so you can make it whatever shape you need and build it yourself cheaply. Learn more->

[Jaehwi Song @j__aehwi]

Introducing HABIT — a large-scale robot manipulation dataset for human-present environments, where a person shares the workspace and interacts with the robot in every episode.

60 tasks · 10,563 episodes · 164 hours of rich human-robot interaction.

Toward robots that are not just capable, but safe and socially compatible around people.
https://t.co/kEtkqbuoIn

🧵[1/7]

[Saquib Alam @saquiboye]

@nezbuilds submitted

[Saquib Alam @saquiboye]

@nezbuilds Sure, thank for this

[Saquib Alam @saquiboye]

Step closer towards creating a better reward functions than simplify imitating human action.

Imitation has no learning feedback loop and doesn’t allow for improvement. Instead using the video to guide the motions as well as reward or penalise based on the action creates that essential feedback loop for improvement. Interesting research!

[Hanzhi Chen @hanzhic678]

🤖🎥 We have recently seen some cool works that leverage human videos to learn robot policies, even without robot demonstrations.

But what if human videos could do more than teach robots what to imitate?

We show that human videos can teach robots predictive representations of action, dynamics, and value. These embodiment-agnostic representations transfer across robot embodiments, enabling robots to self-improve from their own rollouts and failures - without online human intervention. Introducing:

📄 Robot Self-Improvement via Human-Video Dynamics Models

Our method enables two different robots to self-improve across 7 real-world manipulation tasks:

🚀 40% → 81% success rate
✅ zero human intervention during improvement
🌈 works across robot embodiments and different policy backbones

Human videos are not just data for imitation; they can support robot self-improvement.

🧵👇

[Saquib Alam @saquiboye]

Robots getting closer to human sensory system.

[Sharpa @SharpaRobotics]

What if the next leap in robot manipulation comes from touch, not just vision?

To get there, foundation models need to understand tactile feedback the way they understand images and language. And tactile policies cannot be locked to specific hardware (that makes real-world deployments & maintenance quite complicated).

FTP-1 solves both. One of the 1st foundation model for touch. 21 sensors. ~3,000 hours of data. Transfers to hardware it has never seen before.

+17% on known hardware. +31% on never-seen hardware.

We're proud this research led by @michaelyuancb ran on #SharpaNorth, #SharpaWave hands, and our DTC sensors.
Special thanks to the teams at @Tsinghua_Uni , @UCBerkeley , @ETH , and @sjtu1896.

Project page: https://t.co/07BckCfoPj

[Saquib Alam @saquiboye]

@sridharfyi @16vchq Vibe video promotional video for your product using Claude - https://t.co/6zmvqWUJil

[Saquib Alam @saquiboye]

@foxtomb232 Create promotional videos for your product using TheFluxTrain and Claude

[Saquib Alam @saquiboye]

@christophersaum https://t.co/6zmvqWUJil - vibe your promotional product video

[Saquib Alam @saquiboye]

@Guronnimo https://t.co/6zmvqWUJil - vibe you promotional product video in Claude. No complicated video editing

[Saquib Alam @saquiboye]

@sajibrahman206 @Guronnimo @JiroBuild I built a product to vibe video a promotional product video using Claude

[Saquib Alam @saquiboye]

@Guronnimo @maybe_mehdi Use Claude and TheFluxTrain to create awesome product videos

[Saquib Alam @saquiboye]

@lightsilver323 https://t.co/6zmvqWUJil - shortfilm and promotional video using Claude

[Saquib Alam @saquiboye]

@DanielSmidstrup Claude makes shortfilm

[Saquib Alam @saquiboye]

@sridharfyi Netflix and chill

[Saquib Alam @saquiboye]

@16vchq https://t.co/6zmvqWUJil - All in one AI filmmaking using Claude

[Saquib Alam @saquiboye]

@Joi2James https://t.co/eFdbe6SiTy