DESKTOP ENTERPRISE TERMINAL • 800+ → 4 SETUPS • 5-ARCHETYPE SYSTEM • SHIPS DAILY • $18/MO

3PM Daily: Closing Auction Debrief in 5 Minutes

Traders drown in 800+ tickers at the 3PM close. They don't need another scanner — they need a definitive debrief they can act on in 5 minutes. 3PM Daily scans 800+ equities in 4.2s with deterministic TA-Lib, surfaces 4 high-conviction setups with entry/stop/target, and ships daily with zero hallucinated prices.

0.00 CLS • 100/100 CWVTA-Lib deterministic • No LLM price guessDesktop-first • Zero dead space
View Live at the3pmdaily.com →
800+ scanned • 4 setups • Ships <30 min/day
Project Meta • Gen-Pop
RoleMulti-disciplinary: Research → Design System → Code
ScopeDiscovery → Shipped • Ships Daily
StackNext.js 16, Tailwind JIT, TA-Lib, Python, Vercel
Infra$18/mo • No GPU • Python daemon
Outcome4 setups in 5 min • Live
Why This Matters to Any Recruiter / Hiring Manager
This is enterprise workflow simplification — same skills that simplify complex processes for non-technical teams in any industry. Dense data tables, zero dead space, WCAG AA, deterministic math where LLMs shouldn't guess. Ships daily.
01 / DISCOVERY

Traders Don't Need More Data. They Need a Decision.

Problem • 800+ Tickers at 3PM
Drowning in scanners, no time for synthesis.
V1 logs showed analysis paralysis at close — 800+ equities, 15 tabs, no definitive brief. Need 4 high-conviction setups, not 800 watchlist items.
Artifact: 11 trader interviews + V1 logs
Insight • 5 Jobs by Time Budget
Not one persona, 5 jobs in 5 minutes.
Mapped jobs: Scalper (30s), Risk Manager (60s), System Analyst (2m), Momentum (3m), Daylight Reviewer (5m). Each needs different density and signal.
Artifact: Time-budget job map
02 / RESEARCH

8 Widgets → 5 Workspaces, Zero Dead Space

Workflow → Widgets → Workspaces
What I Did
Designed 8 dense widgets (debrief, scanner, levels, tape, setup, risk, session, archive) that assemble into 5 workspaces by time budget. Desktop-first enterprise terminals at 1440px, not responsive afterthoughts.
Workflow MappingDesktop-First 1440pxZero Dead Space
Widget List
1. Debrief Header — 5-min narrative
2. Scanner Matrix — 800+ → 4
3. Level Grid — entry/stop/target
4. Tape — closing auction
5. Setup Card — high-conviction
6. Risk Drawer — RR, size
7. Session Clock — 3PM countdown
8. Archive — previous debriefs
03 / SYSTEMS

One Inference Engine, One State Machine — One Core Architecture

5-Archetype Token System • Lives in 3PM Daily • Also in /lab & /work/design-system
PersonaArchetypeBackgroundJob
Night Owl ScalperTokyo Midnight#0A0E1A30s scalp
System AnalystDARPA Radar#121A142m system
Risk ManagerAmber 1981#1A160F60s risk
Momentum TraderCME Synthwave#0F0A1A3m momentum
Daylight ReviewerGeneva Swiss#FAF9F65m debrief
Space-separated RGB tokens: --bg: 10 14 26 → rgb(var(--bg)). <1ms swap, 0.00 CLS. One codebase, 5 cultures. Now also explorable at /work/design-system?product=3pm-daily&persona=*.
04 / CRAFT

Enterprise Doesn't Mean Boring

Dense by intent: Zero dead space, data tables with Mono 12/14 for prices, Inter 13.5 for UI, Instrument Serif for narrative. WCAG AA 4.5:1 on all data.
Status colors semantic: #FF4D23 only for signal, never decoration. Green/red only for entry/stop/target — never for branding.
1440px design: Designed at 1440px for desktop-first terminals, not mobile-first responsive afterthought. Traders live in multi-monitor.
Type scale disciplined: Instrument Serif for debrief narrative, JetBrains Mono for prices and meta, Inter for controls.
05 / AI LEVERAGE

Math Does Math. LLMs Do Context.

Rule — Never Let LLM Hallucinate Price
Deterministic TA-Lib for stops/targets, Python for scan, LLMs only for debrief narrative. Same principle as Artstache GraphRAG grounding.
Guardrail
TA-Lib: entry, stop, target = math
LLM: narrative, context = language
Never cross streams.
Ops — Python Daemon, $18/mo
Next.js 16 + Vercel + Python daemon scans at 3PM ET daily, builds static debrief, ships in <30 min/day. No GPU, no M0 overage, 100% uptime via self-healing.
Cursor + ClaudeTA-LibPython Daemon
06 / OUTCOME

Shipped Daily, No Bloat, No Bullshit.

Live at the3pmdaily.com, ships every market close. Not a concept — production terminal used daily.

Design Outcome
8 widgets → 5 workspaces by time budget, zero dead space, 5-archetype token system from one codebase, WCAG AA, desktop-first 1440px.
Technical Outcome
800+ equities scanned in 4.2s via TA-Lib, deterministic entry/stop/target, Python daemon, $18/mo infra, <30 min/day ops, 100% uptime.
What I'd Measure Next
Time to decision (target <5 min), trust in levels (do traders act on them?), reduction in tab switching, qualitative feedback on whether debrief replaces manual scan.
Outcome — Verified in Prod, Not Vanity Metrics
● Live • Ships Daily
800+ → 4
Scanned to setups
4.2s TA-Lib, deterministic.
5 min
Debrief
From noise to decision.
$18/mo
Infra
Vercel + Python, no GPU, no M0.
Transferable pattern: Complex workflow → time-budget jobs → widgets → workspaces → token-governed design system → deterministic guardrails → ships daily. Same architecture works for any enterprise terminal, logistics dashboard, or creative tool where operators need decisions, not data.
Let's Talk • Available in Duarte, CA

Building dense tools for non-technical teams under time pressure?

I'm a multi-disciplinary designer — Product, Brand, Design Systems, Code. I partner across the stack from discovery to live production. Based in Duarte, CA (91010) — available for hybrid in Glendale, remote, or on-site. 0-2 weeks notice.

What I bring
• Product, Brand, Design Systems, Code
• Figma → Next.js 16 → Prod
• TA-Lib, Neo4j, GraphRAG
• 0-to-1 to scale, 10+ years
the3pmdaily.com • #FAF9F6 / #141413 / #FF4D23 • 0.00 CLSBuilt with Next.js 16, Tailwind JIT
Next Case Study • Configurable
Meta-Portfolio • 82% → 12%0.00 CLS • 100/100 CWV

don-the-imaginator.com: Reserved Landing + Lab

From 82% bounce to 12% — 5 JIT themes with <1ms swap, 0.00 CLS, two-tier AI, self-healing daemons. Then chose restraint.

© Don Martirez — 3PM Daily Gen-Pop • Multi-disciplinary • #FAF9F6 / #141413 / #FF4D23 • 0.00 CLS • Ships Daily