Question: where would a flash-resident tiny language runtime fit relative to TinyMaix-style MCU inference?

[Alpha-Guardian]

Hi TinyMaix folks,

I wanted to share a small MCU language-runtime experiment and ask whether systems like this feel adjacent to the kind of tiny inference TinyMaix represents.

We built a public demo line called Engram and deployed it on a commodity ESP32-C3.

Current public numbers:

• Host-side benchmark capability

  • LogiQA = 0.392523
  • IFEval = 0.780037

• Published board proof

  • LogiQA 642 = 249 / 642 = 0.3878504672897196
  • host_full_match = 642 / 642
  • runtime artifact size = 1,380,771 bytes

Important scope note:

This is not presented as unrestricted open-input native LLM generation on MCU.

The board-side path is closer to a flash-resident, table-driven runtime with:

• packed token weights
• hashed lookup structures
• fixed compiled probe batches
• streaming fold / checksum style execution over precompiled structures

So this is not a standard tiny dense model path. It is closer to a task-specialized language runtime whose behavior has been pushed into a compact lookup-heavy execution form.

Repo:
https://github.com/Alpha-Guardian/Engram

I’d be curious whether you would see this as:

• adjacent to TinyML-style MCU inference
• a different compression endpoint
• or a separate class of language-runtime specialization