Portfolio · 2026

Alienware projects, leveraging AI.

One reusable, no-AI engine (Project 01) feeds three apps. Each adds AI only where it pays off, sized to how hard the task is.

— up · checked —
Project01✦
scrapers-libVisit ↗
The EngineNo AI
The shared toolkit that collects the raw data every app is built on. Plain code, written once, reused everywhere, no AI involved.
🌐 Fetches web pages🧹 Cleans & extracts📐 Standardizes format🐍 Python library
SourcesReddit · News
YouTube · Product pages
Click to see how it works ↕
scrapers-lib · how the engine works, all rules, no AI
STEP 01Fetch pages from the web
→
STEP 02Clean & extract the useful bits
→
STEP 03Standardize into one common format
→
STEP 04Hand clean data to Projects 02-04
PROJECT 02No AI

competitive-database

Visit ↗

A searchable database of competitor gaming-laptop specs.

No AI at all: every value matched by code, checked by a person.
RunsOn-demand refresh
Cost to run$0 / mo
Rules / code
Flip for steps ↻
How it works
Read the vendor's spec page
via the engine, Project 01
Rules
Pull out each spec value
regex & unit parsers
Rules
Merge data & flag conflicts
combine sources, catch disagreements
Rules
Human reviews flagged items
nothing is ever guessed
Human
Deterministic by design. Proof that not every problem needs AI.
PROJECT 03Moderate AI

pulse-check

Visit ↗

Voice-of-customer briefs, what people say about each product.

AI is one part; the rest is plain code: ingest, tally, storage.
RunsIngests daily · Analyzes quarterly
Cost to run~$5-8 / mo
Qwen 2.5 7BHaikuSonnet
Flip for steps ↻
How it works
Sort posts: reviews vs deals
keeps ad roundups out
Qwen 2.5 7B
Tag each comment's topics & tone
the high-volume step
Qwen 2.5 7B
Group repeated opinions
so quotes aren't shown twice
Haiku
Write the product briefs
the harder reasoning task
Sonnet
Free local Qwen 2.5 7B does the high-volume sorting & tagging; Haiku groups duplicates; Sonnet does the writing.
PROJECT 04Heavy AI · cascade

gaming-chatter

Visit ↗

A weekly gaming-news digest & live dashboard.

AI is one part; the rest is plain code: ingest, grouping, dashboard.
RunsIngests daily · Analyzes weekly
Cost to run~$8-12 / mo
Local · nomicHaikuSonnetOpus
Flip for steps ↻
How it works
Read & label each story
summarize, tag, score
Haiku
Group similar stories
local embeddings + numpy matching
nomic-embed
Name each story group
one headline per cluster
Sonnet
Write & fact-check weekly summary
drafts, then reviews itself (×2)
Opus
The full ladder in one app, free local embeddings (nomic-embed-text) for the bulk, Opus twice for the single hardest task.
Click the highlighted rung again to clear ✕