Juan Pablo Gallegos Work H2HI

H2HI

Pulling structured insight out of raw audio — back when AI in a product wasn't the default.

Full-Stack Engineer 2021 React Python Speech-to-text NLP
Role
Full-Stack Engineer
Duration
~8 months
Stack
React · Python · Speech-to-text
Context
Conversation-intelligence tool
Proof summary

What changed before the long read

The case is about turning raw audio into business-readable output when AI product patterns were still immature.

Pipeline
Recorded calls became transcripts, extracted signals, and readable summaries.
Product layer
The UI showed insight instead of exposing model confidence and internals.
Ownership
A small team kept React, Python, and speech-to-text moving as one system.
01

Context

H2HI listens to recorded business calls and turns them into structured, client-specific insight — the themes, signals, and summaries a team can act on instead of re-listening to hours of audio.

It was a full-stack product with a real job: compress conversations into something a business could actually read.

02

Problem

Raw audio is the least structured data there is. Getting from a recording to "here is what mattered, for this client" means transcribing reliably and then making sense of messy, domain-specific transcripts.

And this was 2021 — before pulling AI into a product was a well-trodden path. There was no obvious playbook to copy.

03

Constraints

Speech-to-text of the era was imperfect, especially on domain jargon and crosstalk, which set a hard ceiling on the input.

The output had to be readable by businesspeople, not data scientists — no confidence scores, no model jargon.

A small team with full, end-to-end ownership of the stack.

04

Role

I was a full-stack engineer across the whole pipeline: the React interface, the Python backend, the speech-to-text integration, and the post-processing that turned model output into something a client could read.

05

System shape

H2HI is a pipeline that takes audio in one end and hands readable insight out the other:

Audio upload
Transcription
NLP post-processing
Extraction
React layer

Audio is transcribed by a speech-to-text service, then cleaned, segmented, and mined for client-specific signal in a post-processing stage. The result is structured insight rendered in a React presentation layer.

Because the front of the pipe is messy, most of the engineering lived downstream — in shaping imperfect transcripts into something coherent.

Insight pipeline
Messy input to readable output
Input
Treat transcription as raw material. Speech-to-text output starts the system; it is not the product.
Processing
Extract client-specific signal. Cleaning, segmentation, and NLP shape messy transcripts into useful structure.
Interface
Hide the machinery. Business users see insight and summaries, not model internals.
06

Key decisions

Decision 01
Treat model output as raw material

The transcription was never the answer; it was the input. The product's value lived in the post-processing that made messy transcripts legible.

Decision 02
Client-specific extraction

Generic summaries were noise. Insight had to be framed in each client's own terms to be worth reading.

Decision 03
Hide the machinery

The presentation layer shows insight, not confidence scores or model internals — businesspeople see conclusions, not plumbing.

07

Trade-offs

Gained

  • Hours of audio compressed into readable insight
  • An AI-in-product capability years before it was common
  • Full-stack ownership kept the pipeline coherent
  • A presentation layer business users could actually act on

Cost

  • Imperfect transcription set a ceiling on quality
  • Heavy post-processing was needed to compensate
  • Bespoke, client-specific extraction was hard to generalize
08

Outcome

H2HI shipped as a working full-stack product that turned recorded conversations into structured, client-specific insight. In 2021 it was also a bet — that AI belonged inside the product rather than bolted on afterward.

09

Lessons

1
When the input is messy, the product is the post-processing. The model gets you raw material; the engineering makes it useful.
2
"AI in the product" is a presentation problem as much as a modeling one — what you hide matters as much as what you compute.
3
Owning the whole stack is what let a small team move a fragile pipeline end to end without losing the thread.