At OrbiSky, we believe in a structured yet flexible approach to AI development. Today, we're lifting the lid on our AI development pipeline, showcasing how we tackle complex challenges, from initial problem definition to real-world deployment. #AI #SoftwareDevelopment #ProblemSolving
About us
OrbiSky Systems is a stealth-mode startup, backed by world-class AI experts, developing a groundbreaking platform with the potential to transform industries. Our focus is on secure, edge-based AI that enables unprecedented autonomy and decision-making capabilities in demanding environments. We're quietly building a future where AI not only empowers machines but also redefines how we tackle complex challenges. Stay tuned for updates as we emerge from stealth mode.
- Website
-
https://orbiSky.com/
External link for OrbiSky Systems
- Industry
- Technology, Information and Internet
- Company size
- 2-10 employees
- Headquarters
- London
- Type
- Privately Held
- Founded
- 2023
- Specialties
- Sustainable Technology, Environmental Impact Reduction, Positive Social & Environmental Change, Artificial Intelligence, Machine Learning, Data Security, Cybersecurity, Supercomputing, and QuantumML
Locations
-
Primary
Get directions
London, GB
-
Get directions
Houston, Texas, US
Updates
-
We're delighted to be part of the Cosmic Capital Programme! Thank you SETsquared Surrey, UK Space Agency, University of Surrey, and partners for this fantastic opportunity. We're excited about the tailored support and networking opportunities to accelerate our growth alongside other innovative space tech companies. Looking forward to the journey ahead! #spacetech #space #innovation 🚀 Ian James, Seraphim Space, Rymdkapital, Space Capital
We are delighted to announce that five innovative space tech companies have recently joined the Cosmic Capital Programme. These startups are making strides in both upstream and downstream space industries across the UK. The Cosmic Capital programme is backed by the UK Space Agency and the University of Surrey, and in partnership with SETsquared Partnership and Space South Central. Our mission is to equip these space businesses with the necessary tools, knowledge, and contacts to secure investment and accelerate their growth. Allow us to introduce Satraka Ltd, Applied Atomics, High Temperature Material Systems (HTMS), Universal Atmosphere Processing (UAP) and OrbiSky Systems. Our program is dedicated to helping companies reach new milestones, and we’ll be instrumental in driving their investment readiness in the next few months with tailored business support. Let’s celebrate innovation, collaboration, and the infinite potential of space exploration! Applications are still open; apply now for an opportunity to join other space-tech companies in the Programme. For further details, visit https://lnkd.in/eZPf_qFj #CosmicCapital #BusinessIncubator #UKSpace #Set2 #Investment #SpaceSouthCentral #SurreyResearchPark #UKSpaceAgency
-
-
Scaling Edge AI efficiently remains a significant challenge that we at OrbiSky are tackling head-on. Effective scaling allows AI applications to handle more complex tasks and larger datasets without performance bottlenecks. Engineers, what scalability issues have you encountered in your work with Edge AI, and how have you addressed them?
-
We've read Apple's paper on OpenELM, and here's what we think: 1. It’s smart to scale layers like they do, but they need a hard comparison with the latest models like GPT-3 or BERT across multiple benchmarks to show it’s not just hype. 2. They say it needs fewer tokens to train—great. But how much compute power are we talking? They should spell out the GPU hours and hardware specs clearly. 3. They’ve open-sourced everything, which is gold for reproducibility. Adding real-world use cases from the community could drive home the model's utility. 4. Including Apple conversion is a cool niche. They should dive into how well it performs on Apple gear in terms of speed and accuracy. 5. The foundation is set; now, tell us where to go from here. They could suggest specific architectural tweaks or new training methods to overcome current downsides. Check out the OpenELM paper and share your thoughts! What improvements or applications do you see? Let’s collaborate and drive this forward.