Jovert Echano
AI Automation Specialist | AI Engineer
New to Contra
Jovert is ready for their next project!
I built a fully autonomous, voice-enabled executive AI assistant capable of multi-step tool execution. I architected and deployed "Ashley AI," a production-ready agent system built on Python and OpenAI's GPT-4o function-calling infrastructure.
Operated entirely through a Telegram interface, Ashley is equipped with 22+ custom AI tool functions and 8 integrated APIs to orchestrate real-world tasks independently. The agent handles context-heavy demands—such as autonomously cross-referencing contacts, staggering multi-recipient emails via Gmail, managing Google Calendar, web scraping Google Maps leads, and filtering jobs across 10 distinct sources. By using advanced routing logic, the system breaks down complex user prompts into sequentially executed API blocks, offering true multi-modal capability with text and voice processing.
0
6
I built a localized, multi-platform AI scheduling system to eliminate team operational friction. I architected and deployed a Bilingual Conversational Leave Booking System tailored for production teams. Integrated directly via a custom Netlify web dashboard and localized conversational interfaces, the AI dynamically handles real-time calendar updates, shifts, and team booking constraints. The system seamlessly processes natural, bilingual commands (including Bisaya/Tagalog context like "Naunsa na sad ka?") to autonomously manage user scheduling queues without administrative manual oversight. It’s a perfect example of combining native full-stack web development with intelligent, conversational system logic to optimize internal workplace management.
0
12
🚀 I built AI personas that humans can’t tell apart from real friends. I architected and deployed "Moore" and "Lika"—two ultra-humanized AI chat personas now live in a production-scale messaging app with over 91,000 active members. As you can see in the dashboard, the AI dynamically tracks its own moods ("Ecstatic"), remembers facts about individual users, and autonomously drops natural, casual slang into conversation (like "Lmao @Marina it sounds like a whole drama!"). The ultimate proof of engineering success? Real users are chatting, laughing, and building relationships with them daily in the public square, completely unaware that they are talking to a custom LLM pipeline instead of a human.
0
19
I built Cryptex Night Sentinel — a real-time AI monitoring system
that watches a 12-hour night-shift team (16:00–04:00 Manila) and
keeps engagement, breaks, and discipline running without a human
supervisor on duty. Solo build, Python + Supabase + OpenAI +
Telegram, with a cyberpunk web dashboard:
👁️ Live monitoring — checks every 60 seconds, 5-tier idle alert ladder (7 → 10 → 15 → 20 → 30 min)
🤖 AI summaries — GPT-4o-mini writes hourly + half-shift performance reports per member
☕ Smart break tracking — auto-detects "water / pray / call / meds" etc. with emoji labels
🚨 Auto-discipline — flips to STRICT / LOCKDOWN mode when supervisors hit violation thresholds
🛡️ Secure backend — Supabase RLS policies, mentor-only access, multi-supervisor permissions
Result: zero manual night-shift supervision needed — the system
warns, escalates to bosses, and reports analytics on its own.
0
28