Ownership-driven product engineer
I build products that hold up after launch.
I take a product from unclear problem to dependable production system: user experience, frontend, backend, infrastructure, and practical AI automation.
I work across the whole product, without losing the product.
What I am into
Modern frontend
Interfaces people can understand quickly.
System architecture
Backends that remain clear under real use.
AI workflows
LLMs used where they create concrete leverage.
Cloud + delivery
Deployments and operations built into the product.
Current Build
Yaar WhatsApp Sourcing Workflow
A WhatsApp-first B2B sourcing backend that turns messy buyer requirements into structured inquiries, routes them to relevant suppliers, captures quotes, and sends buyers a clean comparison.
Runtime Path
From message to matched quote
WhatsApp Webhook
FastAPI Shell
Redis / PubSub
Workflow Orchestrator
Gemini + Rules
Supplier Quotes
WhatsApp-first UX
Buyer and supplier flows run inside WhatsApp: hi/help, onboarding, feedback, inquiry intake, and quote replies.
Deterministic workflow core
A stateful orchestrator controls buyer onboarding, missing specs, supplier routing, quote parsing, and buyer summaries.
Bounded AI integration
Gemini is used for extraction, OCR/audio interpretation, quote parsing, and summaries while business transitions stay explicit.
Cloud-ready architecture
Local Redis/RQ workers can be swapped for Google Pub/Sub and Cloud Tasks in production deployments.
Operational data model
PostgreSQL stores users, buyers, suppliers, inquiries, supplier routes, quotes, messages, audit logs, and credits.
Human-in-the-loop controls
Admin APIs support supplier approval, inquiry correction, message review, quote edits, and billing controls.
Projects Implementations
[ Project ]
Yaar WhatsApp Workflow
Building a WhatsApp-first industrial sourcing product. Designed the user flows, admin review surface, FastAPI backend, stateful workflow orchestrator, supplier routing, quote parsing, credits, and cloud-ready worker architecture.
[ Project ]
BetOnChat
Solving the trust bottleneck in prediction markets. Architected a trustless protocol on EVM that handles real-time staking. Resulted in a 100% verifiable betting mechanism without centralized intermediaries.
[ Project ]
Farad
Eliminating payment latency for digital creators. Engineered a real-time streaming payment engine using Superfluid. Reduced brand-to-influencer settlement time from 30 days to literal seconds.
[ Project ]
Show Off
Bridging the gap between digital assets and physical reality. Built an immersive AR interaction layer for NFTs. Enabled real-time 3D spatial rendering of cross-chain assets on mobile devices.