kensat — What is it?

KENSAT is a CubeSat that runs a large language model on-orbit, addressing the challenge of autonomous AI compute in space.

⭐ 77 Stars 🍴 11 Forks Python NOASSERTION Author: kenchangh
Source: Description per README View on GitHub →

Why it matters

KENSAT is attracting attention due to its innovative application of edge AI in space, addressing the unique challenges of power, thermal management, and reliability in a small satellite environment. Its use of a quantized LLM and custom UHF radio link is particularly unique.

Source: Synthesis of README and project traits

Core Features

Edge AI Compute

Executes neural-network inference on an NVIDIA Jetson Orin Nano within a 2U CubeSat, demonstrating the feasibility of AI compute in space.

Source: README
Power Management

Power-gates the Jetson Orin Nano to conserve energy and manages its operation within a defined energy budget.

Source: README
Custom UHF Radio Link

Utilizes a custom UHF radio link for downlinking inference results to amateur ground stations, making the data accessible to a wide audience.

Source: README

Architecture

The architecture is modular, with distinct subsystems such as the OBC, radio, EPS, burnwire, jetson, and ground-station. It employs a fault-tolerant design with one-shot state machines for critical operations and signal-integrity-aware bus design for communication.

Source: Code tree + README

Project Knowledge Graph

Knowledge graph: project (center) + core features (inner hexagons) + key dependencies (outer chips) NVIDIA Jetson Orin NanoNVIDIA Jetson… Si4463 UHF transceiverSi4463 UHF tra… AS179 T/R switchAS179 T/R swit… Edge AI Compute Power Management Custom UHF Radio LinkCustom UHF Radio Li… kensat Project Core feature Key dependency

Center: project; inner ring: core feature modules; outer ring: key dependencies. Auto-generated from core_features and tech_stack.key_deps.

Tech Stack

LanguagePythonFrameworkSalvo RTOS for OBC, KiCad for PCB design, scikit-rf for RF design
NVIDIA Jetson Orin NanoSi4463 UHF transceiverAS179 T/R switch
Proprietary vendor SDKs for OBC firmware
Source: README, Code tree

Quick Start

Not enough information.
Source: README Installation/Quick Start

Use Cases

KENSAT is suitable for researchers and developers interested in edge AI, space technology, and amateur radio. It is useful for demonstrating the capabilities of AI in space, developing new communication protocols, and contributing to the amateur radio community.

Source: README

Strengths & Limitations

Strengths

  • Strength 1: Unique application of AI in space
  • Strength 2: Open-source and accessible to the amateur radio community
  • Strength 3: Modular and fault-tolerant design

Limitations

  • Limitation 1: Proprietary SDKs for OBC firmware
  • Limitation 2: Limited documentation on building and running the system
Source: Synthesis of README, code structure and dependencies

Latest Release

Not enough information.

Source: GitHub Releases

Verdict

KENSAT is a pioneering project that showcases the potential of AI in space. It is of interest to those working at the intersection of AI, space technology, and amateur radio, offering a unique platform for innovation and collaboration.

Frequently Asked Questions

What is kensat?

KENSAT is a CubeSat that runs a large language model on-orbit, addressing the challenge of autonomous AI compute in space.

What are the main features of kensat?

kensat's core features include: Edge AI Compute, Power Management, Custom UHF Radio Link.

Why is kensat trending?

KENSAT is attracting attention due to its innovative application of edge AI in space, addressing the unique challenges of power, thermal management, and reliability in a small satellite environment.

What is kensat used for?

KENSAT is suitable for researchers and developers interested in edge AI, space technology, and amateur radio.

Transparency Notice
This page is auto-generated by AI (a large language model) from the following public materials: GitHub README, code tree, dependency files and release notes. Analyzed at: 2026-07-12 18:33. Quality score: 80/100.

Data sources: README, GitHub API, dependency files