Zacky Muhammad Dinata@zmdinata
2
Events attended
2
Submissions made
Indonesia
1 year of experience
About me
I am an Information Systems student at STMIK IKMI Cirebon with one clear destination: building intelligent systems as an AI Engineer. My journey here is anything but conventional. It started in 2021 as a freelance digital illustrator, completing over 60 artworks for global clients across the NFT and digital art space. That chapter sharpened my eye for detail, taught me how to communicate across cultures, and gave me a design sensibility that I now bring into every technical project I build. From there, I moved into Web3 research, data analytics, and eventually machine learning. Each step was intentional. Each skill is built on the last. Today I work with Python, Scikit-learn, TensorFlow, Keras, Pandas, and the full data science workflow from exploratory analysis to model evaluation. I am currently developing an ML research paper targeting SINTA 2+ publication standards, studying the behavioral impact of social media through a machine learning lens. In early 2026, I was selected as a scholarship recipient for the Pijak x IBM SkillsBuild AI Engineer Intensive Program, one of the most structured AI development programs available in Indonesia today. The program covers everything from Python fundamentals and Git to Deep Learning and end-to-end Machine Learning system development, backed by IBM. I believe the best engineers are not just technically sharp. They communicate clearly, think critically, and build things that actually matter to real people. If you are working on anything in machine learning, data, Web3, or creative tech. let's connect! Portfolio: zmdinata.web.id
🤝 Top Collaborators
.png&w=640&q=75)
🤓 Latest Submissions
AutoOverlay AI: Autonomous Options Yield Agent
Overview AutoOverlay AI is an autonomous options-income overlay agent built for the Alpaca AI Trading Agents Hackathon. Managing a $100,000 USD Alpaca Paper Trading portfolio, it monetizes idle equity via automated Covered Calls and Cash-Secured Puts while rigorously hedging tail-risk. Key Innovations 1. 6-Persona Investment Council: Evaluates every candidate before capital deployment. Personas like Benjamin Graham (enforcing Ch. 14 defensive rules & Ch. 20 Margin of Safety), Warren Buffett, and Ray Dalio debate trade sentiment and log full reasoning traces. 2. Monte Carlo Merton Jump Risk Engine: Stress-tests 1,000 multi-day portfolio paths under non-Gaussian shock jumps (lambda=0.05, jump_mu=-5%) to measure 95% Value-at-Risk (VaR) and Sortino downside ratios. 3. Zero-Slippage Execution & Kill-Switch: Eliminates the 3-10% premium drag of market orders via strict OCC option contract resolution and mid-price limit orders ((Bid+Ask)/2). Includes a Layer 1 Pre-Trade Kill-Switch (>5% drawdown or 3 consecutive stop-losses). Architecture & Team - Quant & AI Layer (Zacky - AI Lead): Monte Carlo jump diffusion, multi-agent council logic, OCC resolver, and risk mitigation. - Backend & Bot Integration (Aji): High-throughput FastAPI services, Alpaca API event loops, and cloud bot deployment. - Frontend & Infrastructure (Aditya): Next.js App Router telemetry dashboard, real-time committee UI, and dedicated VPS hosting. Result Tested on live $100k Alpaca paper capital, AutoOverlay AI successfully filtered illiquid contracts, debated optimal strikes, and executed hedged limit orders—yielding steady overlay income with zero slippage and robust capital preservation.
4 Sep 2026
A2Z Agentz: Autonomous Web3 Scavenger
A2Z Agentz is a fully autonomous, dual-brain AI agent system purpose-built for the AMD-powered Web3 stack, targeting the Base network. In crypto, speed and accurate intelligence are everything — so we split the workload across two specialized agents to maximize both reasoning depth and execution speed. Agent A (The Scout) is the intelligence hub. It runs Meta-Llama-3.1-8B-Instruct-AWQ-INT4 locally on AMD Instinct MI300X via vLLM + ROCm, exposed securely through a Cloudflare tunnel. Agent A performs deep OSINT signal detection — continuously scraping and analyzing DexScreener and Farcaster (Neynar) to score market opportunities in real-time using strict JSON-mode reasoning, keeping end-to-end latency under 25 seconds. Disclaimer on Technology Tags: Please note that while Agent B is powered by the newly released DeepSeek V4 Pro for superior execution logic, the platform's technology dropdown only offered "DeepSeek V3" as the closest available option. We selected the V3 tag to reflect our ecosystem choice, while actively utilizing the V4 Pro API in our live production environment. Agent B (The Vault) is the secure execution engine. Connected via a database queue, it receives high-confidence signals from Agent A and uses Fireworks llama-v3p3-70b-instruct to make split-second, security-gated on-chain execution decisions. The system ships a production Vercel dashboard with real-time WebSocket + polling fallback for live agent logs, JWT / API-key authenticated backend, and a manual Circuit Breaker to pause autonomous on-chain execution. Agent B executes real native transactions on Base mainnet with operator-configurable spend caps (hard $2.00/tx autonomous limit) and optional GoPlus security signals. Built entirely on AMD infrastructure for the AMD Lab Lab hackathon — from MI300X inference to Cloudflare tunnel edge delivery.
13 Jul 2026
