darknite

- Architected a multi-sided SaaS platform serving 20 enterprise clients across 20 isolated tenants - built supply-side and demand-side abstractions with hierarchical RBAC, API key provisioning, cross-tenant asset sharing, and role-specific access flows - Designed and shipped a full user onboarding system enabling new users to self-serve from day one - reduced feature-related support queries by 60-80% and accelerated time-to-value across all 20 tenants - Built a multi-step workflow automation engine with rule-based distribution logic matching campaigns to channels - eliminated manual handoffs and consolidated fragmented ops into a single orchestration layer - Implemented agentic AI systems (chat agents, workflow agents, research assistant) with full-stack UIs for creation, management, and real-time execution - automating previously manual processes averaging 10 agent runs per tenant - Enforced TypeScript discipline across a large codebase, maintaining quality at scale; sustained 2 staging builds/day and 1 production deploy/week - shipping fast without sacrificing stability across a live enterprise platform - Tech Stack: Next.js, React, TypeScript, Redux Toolkit, shadcn/ui, Driver.JS, Tailwind CSS, FullCalendar, TipTap, Node.js, REST APIs, OAuth, Multi-tenant SaaS

- Designed and implemented a zero-commission incentive model for a two-sided ride-sharing marketplace - engineered the pricing engine, fare calculation logic, and driver-facing earnings transparency into both the mobile app and backend API - Built the full-stack architecture for both sides of the marketplace: a React Native Rider App with end-to-end ride flow, real-time matching, and payment integrations, backed by Node.js (Express.JS) on AstraDB (managed Cassandra service) - Validated supply-side dynamics through surveys with 50+ active drivers; applied to Y Combinator and incorporated reviewer feedback into platform architecture