Parslet

• 

  Battery Aware Demo

• 

  Parslet Help Menu

• 

  Image Filter Demo

• 

  Hello Demo

Inspiration

As an Outreachy intern with Parsl, I saw how powerful parallel task execution could be. But I also realized how out of reach it was for users like me, Android users without access to clusters, servers, or HPC setups. I wanted to make Parsl’s core ideas available to people with nothing more than a smartphone or tablet.

What it does

Parslet is a lightweight parallel task runner that works in low-resource environments like Android (via Termux). It supports DAG-based workflows, battery-aware scheduling, and offline execution. You can create tasks, connect dependencies, and run the whole thing, without needing a cluster or pip install.

How I built it

I used Python and built a custom DAG engine from scratch, optimizing for low memory and CPU usage. I developed and tested directly on an Android tablet using Termux and open-source tools. All visual assets and docs were designed on my Lenovo android tablet, fully mobile-first.

Challenges I ran into

• No access to high-performance computing gear
  
• Limited electricity and RAM
  
• No pip/venv support; so I had to avoid heavy dependencies
  
• Needed to create visualization and execution logic from scratch
  
• Battery API access was tricky on Android
  

What I learned

• How to think modularly and keep things lightweight
  
• That user empathy should guide dev decisions
  
• How to optimize Python for slow systems
  
• How to document and test entirely without a laptop
  

What's next

• PyPI package submission
  
• Offline GUI for Android
  
• Support for custom task types and plugins
  
• Deeper integration with existing Parsl tools
  
• More examples + native Termux install scripts

Built With

• acode
• android-shell
• bash
• canva
• drawio
• github
• python
• termux