This site is alive
A real terminal, not a video. Type help and press enter. Go on, poke around.
↑↓ history · tab autocomplete · try secret
// quick commands
- : real name + alias
- : best builds
- : what I ship with
- : critical intel
- : find me online
- : wipe the screen
psst… every command on this page runs in your browser, no servers were harmed.
Work gallery
A collection of systems, apps, bots and experiments I’ve shipped in public: newest first, coffee not included.
Catalyst8 Signal
Multimodal market signal engine integrating Gemini multimodal LLM for automated sentiment and technical indicators.
NUSA AI
AI Community Operating System for Indonesian RT/RW communities: automated resident reports and AI insights.
Campuss Blink Pass
Solana Blink Actions check-in flow built for a Solana Indonesia grant project: register, check-in, and claim POAP.
drag / swipe cards · ← → keys · click side cards to fan
What I build with
Web apps, Telegram bots and blockchain experiments, from database-driven interfaces to headless agent tools.
Automation, Bots & Scraping
01
End-to-end Python & Node.js automation suites: high-frequency web scrapers, production Telegram bot APIs, scheduled background cron tasks, REST/WebSocket streams, and ETL JSON pipelines that run 24/7 without manual intervention.
⚡Autonomous AI Agents & Feeds
02Integrating Google Gemini multimodal reasoning into automated decision engines, machine-to-machine x402 HTTP micropayment flows, unattended market surveillance bots, and automated GitHub CI/CD test runners.
🧠 AI Data Quality, Evaluation & Web Systems
03Years of hands-on remote AI engineering work: building automated dataset collection and alert scripts, performing high-precision RLHF prompt ranking, and hallucination audits paired with clean monitoring dashboards.
Latest from the repo
No screenshots: this is live data straight from GitHub. What I pushed last, and how often I ship.
▸ recent commits
● LIVE- 4d ago feat(otomasi): penjadwal otomatis + anti-spam guard ARM
- 4d ago refactor(ui): mapping sistem operasional multi-agent ARM
- 6d ago test(core): benchmark script & automation worker pipeline TES
- 4w ago feat(auth): sign-in & sign-up token security flow NUS
- 4w ago perf(db): migrate high-throughput document database NUS
- 6w ago feat(bot): async automated worker with task queue BLU
github seems shy right now, refresh or check github.com/nialthony ↗
The human behind the terminal
drwxr-xr-x nial dev since-2024
indonesia, gmt+7 · ships 24/7 anyway
DHANIAL ANTHONY ✓
a.k.a Nial Thony, that's what the internet calls me. Handle: @nialthony everywhere.
I’m an Automation Engineer & Python Developer based in Indonesia, specializing in building unattended data-scraping pipelines, production Telegram bots, and autonomous AI systems that eliminate manual workflows.
On the engineering front, my public GitHub features 20+ repositories focused on high-reliability automation: production bot architectures with 94 automated security tests (Paybox Agent Wallet), continuous market scrapers and CRT alert suites (CekRego), multimodal AI signal pipelines (Catalyst8 Signal), and multi-chain CLI minting runners.
Since 2023, I’ve applied this automation-first mindset to remote AI engineering platforms (Outlier, Alignerr), crafting automated data ingestion scripts, conducting RLHF prompt evaluations, and performing rigorous data validation.
My core engineering philosophy: automate the repetitive, verify with precision, and ship reliable systems that run quietly in the background 24/7.
Automating
data pipelines, scraping suites & Telegram botsAutonomous AI Bots
& agent wallet protocols (x402)Async Python
& high-throughput background daemonsKopi susu ☕
and zero-downtime deploysThe journey
Where the work actually happens: remote, freelance, and self-directed since 2023.
Automation Developer & AI Data Specialist
Independent Contractor · Remote Automation & AI Workflows
- Built automated data-collection, web scraping, and alert scripts with asynchronous Python and the Telegram Bot API to eliminate manual data entry and streamline real-time reporting.
- Engineered scheduled data-ingestion pipelines normalizing multi-source JSON feeds with automated error recovery and continuous monitoring.
- Conducted prompt-response ranking, RLHF evaluations, and hallucination auditing on LLM outputs across complex technical domains on Outlier and Alignerr.
- Produced and audited 100+ technical dataset modules achieving 98%+ accuracy against strict engineering style guides.
$ plus continuous self-directed learning in LLM evaluation, audio data QC, prompt engineering & async Python. proof of work ↗
LET’S BUILD
SOMETHING
Got a project, a wild idea, or a spot on your team? My inbox is open, and I reply fastest on Telegram.