Agentic AI Build Week
Back to People

Khidayotullo Salakhitdinov

Entrepreneur (Founder/Co-Founder)

AI experience: 1–2 years

Languages: Python

Public projects: 2

Public projects

Projects this builder chose to show on their profile, linked to the published project gallery.

Elemental Arena

Track: Robotics & Physical AI · Khidayotullo

# ⚔️ Elemental Arena **Every AI you've ever used lives behind glass. Ours sits on your desk — and it wants to fight.** Elemental Arena turns a $20 ESP32-S3 voice device into a game console and a companion in one: you design an AI-generated elemental beast in your browser, it lives on your physical puck as both your fighter and your personal assistant, and two pucks on a table can battle each other head-to-head by voice — no controller, no screen you have to hold. It's for anyone who's ever wanted their AI assistant to *be* something instead of just answering questions — makers, hackathon judges, and anyone who owns (or wants to build) a talking ESP32 device. **Chatbots answer questions. Ours throws fireballs.**

View Project

Track: Retail & Hospitality · PharmKulen

Retail and hospitality run on four manual reads. Someone reads the supplier dossier — retyping it, checking it against rules they remember, chasing missing documents by email. Someone reads competitor prices — occasionally, manually, stale by the afternoon. Someone reads the daily numbers — a morning of spreadsheets before a single rate or inventory decision. And someone reads the reviews — channel by channel, screenshot by screenshot, after the damage is done. Four jobs, one failure: the insight arrives after the moment to act. Backhouse is one commerce platform with four agent desks in the back office: 1. GreenLane turns a supplier's PDF dossier into a cited, rule-checked case — agents classify, extract with citations, and check every claim against rules sourced to named Vietnamese law. Suppliers go live in days, not weeks. 2. GoingRate turns the inventory you already stock into an always-on price radar — a snapshot agent watches competitor channels, classifies every gap, and alerts the same day. One catalog, no second tool. 3. ForwardBook reads the day's reservation data and hands your commercial team a decision, not a spreadsheet — demand forecast by date and segment, with recommended actions and the evidence behind each one. 4. StreetTalk pulls every review, ticket, and social mention into one Voice-of-Customer view — sentiment, share of voice against competitors, alerts on emerging issues, and an analyst agent that answers questions with cited evidence. One rule across all four: agents read, humans decide. Nothing auto-approves a supplier, reprices a shelf, changes a rate, or replies to a customer. Every agent step is logged and auditable — when a regulator, a boss, or a judge asks "why did the system say that?", the answer is on the screen. The back office used to be where time went to die. Now it's agent teams and an audit trail.

View Project