English for Engineers · the hub

English Hub

One personalized path across what we build, what we delegate to your AI, and what we curate.

A2level→ B1

0 days

TODAY

Today's cycle

SRS → input → output

Review0 cards due

15′ left 2

Comprehensible inputread · watch · listen — and log your external hours

45′ left 3

Output · writing

15′ left

01 · INSTRUMENT

Coverage

How much of this corpus you can already read

0^%words known

75% functional90% fluent reading

Corpus · Backend Engineering3,759 word families

A2 · core

B1 · working

B2 · advanced

≥ 90% 75–90% < 75%

frequency coverage, after Nation — 98% needed for unaided reading

Signature. Your known-word profile measured against a frequency corpus — dashed lines mark the 75% and 90% thresholds. An instrument, not a score.

METRIC · HOURS

Input hours

Hours of comprehensible input decide progress

Log what you watched, read or listened to OUTSIDE the site too — that volume is 80% of the gain.

0mtoday

0mthis week

0mtotal

0m active input0m passive0m review0m output

Log external input

02 · ORCHESTRATOR

Next — one path, three modes

Sequenced across tools, not within one

OwnRead “Reading a code review”
Reading at your level locks new words in context — the comprehensible-input sweet spot (i+1).
~10 min·→ Reading (here)·comprehensible input · Krashen
Open reader→
OwnDrill a grammar topic
Targeted pattern practice — items generated on the fly with instant feedback, scheduled like your vocabulary.
~6 min·→ Grammar (here)·explicit practice · DeKeyser
Open grammar→
DelegateExplain a concept aloud — 90 seconds
You can read it; now produce it. Your AI coach plays a skeptical reviewer and probes follow-ups.
~6 min·→ Speaking · your AI (BYOK)·output + feedback · Swain
Open coach→
CurateWatch: “Database indexing, visually — ByteByteGo”
Extensive listening at your level locks in this week's terms. Captions on, watch once for gist.
~12 min·↗ Listening library (external)·comprehensible input · Krashen
Watch↗

03 · PIPE

Bring your own content

Any text → a lesson, five ways

Try:RFC 9293 (TCP)martinfowler.com/microservicesyour last incident postmortem

Stage 01 · ExtractSort every word against what you knowPaste a passage above, then “Make lesson”.

0 known0 new0 technical

Stage 02 · BuildSpaced cards + AI-written exercisesGenerated from the source, ready to drill:

0 SRS cards

Stage 03 · ReuseOne source, five waysEach reuse routes to the mode that does it best:

ComprehensionOwn

VocabOwn

DictationDelegate

RetellDelegate

ImitationDelegate

Or save any phrase by hand:

Signature. One paste becomes a spaced-repetition deck, AI-written exercises, and five reuse modes — the same pipe for an RFC or a news article.

Built here

● Own · live in the product

Vocabulary · spaced repetition

Due today

backend

0cards

No cards yet — start a review or paste content below to build your deck.

spaced retrieval, after Roediger & KarpickeReview now→

Reading · on our own lessons

Continue — Reading a code review

comprehensible · i+1

An annotated reader over the site's engineering lessons. Unknown words are tappable and feed straight into your deck.

0% of this lesson's words known — 6 new terms ahead.

comprehensible input, after KrashenOpen reader→

METRIC · MONOLOGUE

Monthly monologue

Progress you can hear

Due: record ~3 minutes — what you did this month, in English.

No recordings yet — record your point A today.

Routed to your AI

◆ Delegate · powered by your own key (BYOK)

Speaking

Say it out loud

Launchpad

Coach — “The Skeptical Reviewer”

probes assumptions, asks for precision

Structured task we hand off“Explain a concept aloud in 90 seconds, then field three follow-up questions about failure modes.”

fluencytechnical accuracyhedging language

your AI · add your keyOpen in your AI→

Writing — correction

Get it marked up

Launchpad

Coach — “The Precise Editor”

marks register, articles, hedging

Structured task we hand off“Rewrite an incident postmortem for an exec audience. The coach flags article errors, register slips, and weak verbs — with reasons.”

articles a/theregisterconcision

your AI · add your keyOpen in your AI→

Curated, not built

↗ Curate · leveled external library

Listening & immersion · B1–B2 · the engine we deliberately don't build

Database indexing, visually — ByteByteGo

video · 12 minB2intensive: rewatch with captions

How TLS actually works — Computerphile

video · 18 minB1intensive: shadow the narrator

Software Engineering Radio — interviews

audio · 41 minB2extensive: listen once for gist

The Changelog — engineering podcast

audio · 60 minB2extensive: background listening for volume

Hussein Nasser — backend deep dives

video · 20 minB1intensive: pause and mine 5 new terms

Krazam — tech satire (light immersion)

video · 5 minB1extensive: fun input, no pausing

Intensive — study it

Short clips at or just below your level. Rewatch with captions, mine 5–8 new words into your deck.

Extensive — bathe in it

Longer, easier audio for volume. Listen once, no pausing — fluency comes from quantity of input.

From B2 your input should be native — these external sources ARE the curriculum at that level, our texts are the on-ramp.

What we don't build — and where we route youA thin orchestrator beats a bloated app. We own three honest things and point well for the rest.

We don't run a ~~speech-recognition engine~~.→ your AI coach runs speaking practice (BYOK)

We don't ~~grade your writing~~ ourselves.→ your AI marks it up against our rubric

We don't ~~host a video library~~.→ we curate & annotate external sources