English
Minus Zero
Report

Minus Zero

@teamminuszero

Full Stack AI for Autonomous Driving

2 Following1.5K Followers199 Posts
  1. Minus Zero@teamminuszero
    Thrilled to announce a major milestone for our end-to-end driving stack at @teamminuszero — point-to-point autonomous navigation without HD maps using language based prompts!
    Gagandeep Reehal @GagandeepReehal
    Thrilled to announce a major milestone for our end-to-end driving stack at @teamminuszero — point-to-point autonomous navigation without HD maps using language based prompts! With our mapping interface, users can simply enter a destination, and our end-to-end models autonomously drive the full route to get there. Much like a human driver, our system interprets high-level navigation cues — such as “turn left” or “turn right” — from standard mapping services and plans its trajectory accordingly. Since we only rely on high-level guidance, there’s no dependency on dense HD maps, enabling the system to scale rapidly and generalize to new roads. This also means our stack can integrate across multiple mapping solutions with ease. All demo videos below were captured on roads never seen during training 👇 #foundationalmodels #autonomousdriving #fsd #llms
    103413
  2. Gagandeep Reehal✓@GagandeepReehal
    Thrilled to announce a major milestone for our end-to-end driving stack at @teamminuszero — point-to-point autonomous navigation without HD maps using language based prompts! With our mapping interface, users can simply enter a destination, and our end-to-end models autonomously drive the full route to get there. Much like a human driver, our system interprets high-level navigation cues — such as “turn left” or “turn right” — from standard mapping services and plans its trajectory accordingly. Since we only rely on high-level guidance, there’s no dependency on dense HD maps, enabling the system to scale rapidly and generalize to new roads. This also means our stack can integrate across multiple mapping solutions with ease. All demo videos below were captured on roads never seen during training 👇 #foundationalmodels #autonomousdriving #fsd #llms
    Download
    2213675
  3. Minus Zero@teamminuszero
    We're back with another test ride of our Autopilot on Bengaluru roads. Have a look at our end-to-end Foundational Models navigating some complex scenarios like dangerous door-opening in the middle of the road, handling intersections and driving with partially-visible markings.
    Download
    3313862
  5. Minus Zero@teamminuszero
    Testing AVs in rare, risky scenarios is hard & unsafe.
We built an AR simulation layer that overlays virtual obstacles (jaywalkers, cut-ins, roadblocks) onto live camera feeds—enabling real-time edge case testing & faster, safer autonomy development.🚀 #ai #technology #india
    Download
    016425
  7. Vijay Shekhar Sharma✓@vijayshekhar
    🔜🔜👍🏼
    Gagandeep Reehal @GagandeepReehal
    Have a look at how foundational AI model learns to drive a car in busy streets of Bengaluru! 🚀🚀 We are proud to showcase the recent progress we make at @teamminuszero demonstrating end-2-end autopilot system navigating on Indian public roads. Our system gracefully navigates challenging real-world scenarios— unstructured roads and on-coming traffic - be it erratic two-wheelers like Yulu, wandering animals, push carts, or the daily chaos of urban traffic. Yes, it handles cows! Building built and trained indigenously, this is first time someone is demonstrating end-to-end autonomous driving on India roads. And we are already doing pilots with multiple OEMs. Minus Zero has taken an AI-first approach by training end-to-end foundational models for autonomous driving leveraging self-supervised learning from large scale raw data without human labels, instead of traditionally used modular architecture limited by hard-coded rules or pre-defined maps. With demand for ADAS rapidly rising in India, we believe this is the perfect time to introduce a next-gen upgrade for Indian consumers who value luxury, safety, and comfort amid the hustle of Indian roads. We will be sharing more information soon. Follow us and stay tuned. ✨ #foundationalmodels #llms #agents #autonomousdriving #autopilot
    789316.3K
  8. Minus Zero@teamminuszero
    Minus Zero
    RT @GagandeepReehal: Ping me next time you are in India @AravSrinivas Let me take you for a spin in Blr 🚀
    Aravind Srinivas @AravSrinivas
    Impressive autopilot capabilities in Indian streets!
    0680131K
  9. Aravind Srinivas✓@AravSrinivas
    Impressive autopilot capabilities in Indian streets!
    Download
    733183.3K314K
  10. Minus Zero@teamminuszero
    India's first end-to-end autonomous driving was demonstrated on Bengaluru streets — no hard-coded rules or predefined maps, just raw data & foundational models. Read the tech + challenges behind it👇
🔗 
#ai #technology #adas #india #bengaluru
    Download
    10282K
  12. Minus Zero@teamminuszero
    Minus Zero
    RT @GagandeepReehal: 📊 Offline evaluation of e2e models in Autonomous Driving: Why trajectory similarity scores can give a false gratification? Once a model is trained, the next step is evaluating whether it can actually drive — not just mimic past data. A common pitfall here is a lot of research papers over-rely on trajectory similarity score like L2 loss between predicted and actual trajectories, etc. But driving isn’t about perfectly matching one “ground truth” path — it’s about choosing from many valid, safe, and socially acceptable options in a complex world. Similarity scores often fail here because they: (•) penalize valid alternatives, even if they are safe. (•) encourage the model to average across behaviors, producing unsafe, indecisive predictions. (•) doesn’t account for safety, feasibility, or rule compliance. (•) doesn’t reflect closed-loop performance, where small errors compound into big failures. (•) overlook system level issues like sensor noise and latencies which can amplify errors. That’s why at @teamminuszero we are working towards a richer set of evaluation metrics — from multi-modal prediction accuracy to intervention rates, off-road violations, and closed-loop scores.  A robust evaluation framework must capture reliability, generalization, and real-world readiness. #autonomousdriving #fsd #autopilot #llms #foundationalmodels
    0301.9K
  13. Minus Zero@teamminuszero
    Minus Zero
    RT @SriniKay: If the top function of #ADAS & #AutonomousDriving is safety, why can't it be made suitable in a country that loses 173,000 lives annually on its roads? I came across a startup that does just that, by using #AI to adapt to our unique road conditions & behaviour. I write in @autox
    0101
  14. Minus Zero@teamminuszero
    Minus Zero
    RT @GagandeepReehal: It was great chat @SriniKay India and emerging auto markets need better ADAS features and faster adoption - road safety is large enough cause we should be focused on. And there is not better place than India to be the developmental hub for next billion auto users. #adas #autopilot #fsd
    Srinivas Krishnan @SriniKay
    If the top function of #ADAS & #AutonomousDriving is safety, why can't it be made suitable in a country that loses 173,000 lives annually on its roads? I came across a startup that does just that, by using #AI to adapt to our unique road conditions & behaviour. I write in @autox
    0301.8K
  15. Minus Zero@teamminuszero
    Minus Zero
    RT @GagandeepReehal: Recently had the incredible privilege of meeting Indra Nooyi, former CEO and Chairperson of @PepsiCo. The interesting part of the story is that I belong to the generation that used to read about her in high school textbooks 6-7 years back - as a case study in early leadership from Indian diaspora — on what it means to lead with purpose, conviction, and global impact. Meeting her in person just amplified that value. In just a short conversation, I was even more inspired by her leadership demeanor — calm yet commanding, and razor-sharp. It will stay with me for a long time. #Leadership #Inspiration
    020980
  16. Minus Zero@teamminuszero
    Minus Zero
    RT @GagandeepReehal: It was a lovely chat with @brokebiker from @moneycontrolcom talking about @teamminuszero - our vision, the roadmap ahead and how we are building a world class tech in India for the world. Physical AI is poised to be the next disruption in AI, and autonomous driving is the best place to start.
    Arun Padmanabhan @ArunSaysSo
    “Why not India?” @teamminuszero co-founder @GagandeepReehal is building the future of self-driving tech from Bengaluru — not for California highways, but for Indian chaos. His bet? If it works here, it’ll work anywhere.
    0501.2K
  17. Minus Zero@teamminuszero
    Minus Zero
    RT @ArunSaysSo: “Why not India?” @teamminuszero co-founder @GagandeepReehal is building the future of self-driving tech from Bengaluru — not for California highways, but for Indian chaos. His bet? If it works here, it’ll work anywhere.
    0301.8K
  19. Minus Zero@teamminuszero
    🤘🤘
    Gagandeep Reehal @GagandeepReehal
    Have a look at how foundational AI model learns to drive a car in busy streets of Bengaluru! 🚀🚀 We are proud to showcase the recent progress we make at @teamminuszero demonstrating end-2-end autopilot system navigating on Indian public roads. Our system gracefully navigates challenging real-world scenarios— unstructured roads and on-coming traffic - be it erratic two-wheelers like Yulu, wandering animals, push carts, or the daily chaos of urban traffic. Yes, it handles cows! Building built and trained indigenously, this is first time someone is demonstrating end-to-end autonomous driving on India roads. And we are already doing pilots with multiple OEMs. Minus Zero has taken an AI-first approach by training end-to-end foundational models for autonomous driving leveraging self-supervised learning from large scale raw data without human labels, instead of traditionally used modular architecture limited by hard-coded rules or pre-defined maps. With demand for ADAS rapidly rising in India, we believe this is the perfect time to introduce a next-gen upgrade for Indian consumers who value luxury, safety, and comfort amid the hustle of Indian roads. We will be sharing more information soon. Follow us and stay tuned. ✨ #foundationalmodels #llms #agents #autonomousdriving #autopilot
    13232.4K
  20. Gagandeep Reehal✓@GagandeepReehal
    Have a look at how foundational AI model learns to drive a car in busy streets of Bengaluru! 🚀🚀 We are proud to showcase the recent progress we make at @teamminuszero demonstrating end-2-end autopilot system navigating on Indian public roads. Our system gracefully navigates challenging real-world scenarios— unstructured roads and on-coming traffic - be it erratic two-wheelers like Yulu, wandering animals, push carts, or the daily chaos of urban traffic. Yes, it handles cows! Building built and trained indigenously, this is first time someone is demonstrating end-to-end autonomous driving on India roads. And we are already doing pilots with multiple OEMs. Minus Zero has taken an AI-first approach by training end-to-end foundational models for autonomous driving leveraging self-supervised learning from large scale raw data without human labels, instead of traditionally used modular architecture limited by hard-coded rules or pre-defined maps. With demand for ADAS rapidly rising in India, we believe this is the perfect time to introduce a next-gen upgrade for Indian consumers who value luxury, safety, and comfort amid the hustle of Indian roads. We will be sharing more information soon. Follow us and stay tuned. ✨ #foundationalmodels #llms #agents #autonomousdriving #autopilot
    Download
    559648260.5K