Charli Knows is an enterprise project management and delivery platform that brings projects, statements of work, deliverables, feature requests, time, budgets, meetings, proposals, and client communication into one operating system.

The problem it solves is familiar to delivery teams: project context lives in email, spreadsheets, ticketing tools, and chat; estimates are inconsistent; budget health appears too late; and clients need visibility without seeing every internal detail.

The platform embeds an AI assistant, Charli, directly in the workflow. It supports project and deliverable chat, a global intent interpreter, scope analysis, estimates, risk analysis, acceptance criteria generation, SoW planning, time parsing, budget health, invoice summaries, and weekly reporting.

I built the system as a full-stack product with .NET Aspire services, a dedicated streaming chat service, a YARP gateway, PostgreSQL, Entity Framework Core, SvelteKit, TypeScript, TailwindCSS, OpenAPI-generated clients, Hangfire jobs, Azure OpenAI, optional Gemini, Slack and Gmail integrations, MCP-capable agents, and a white-label client portal.

• Projects, SoWs, phases, deliverables, subtasks, dependencies, and acceptance criteria.
• Feature requests with status workflows, screenshots, scope, risk, estimates, and links to deliverables.
• Time entries, budget snapshots, estimate accuracy, burn rate, deadlines, and invoice-ready summaries.
• Agent users with memory, delegated tasks, MCP tool use, and execution history.
• Client users with restricted magic-link access to project data, tickets, deliverables, and limited chat.

The outcome is a single source of truth for delivery, with AI that understands the project, the SoW, the team, the budget, and the client conversation.

Equimeta is a production-oriented equestrian marketplace and operations platform built for high-trust horse transactions, professional services, barn workflows, and buyer discovery. It combines a public SEO-focused marketplace with authenticated workflows for buyers, sellers, trainers, barn managers, affiliates, administrators, and service providers.

I led the end-to-end technical implementation across architecture, backend APIs, frontend product flows, operational tooling, payment and entitlement flows, asynchronous job processing, SEO verification, and Azure deployment hardening.

Product areas built

• Marketplace discovery for listings, wanted posts, saved searches, professional profiles, and recommendations.
• Deal room workflows for offers, messaging, stage progression, purchase documents, deposits, final payment, reviews, and notifications.
• Barn operations tools for managed horses, care logs, medication, feed programs, events, billing, invoices, contracts, and show-day preparation.
• Horse registry features for claim flows, pedigree edits, media, saved horses, genetic tests, comments, and competition summaries.
• Admin tooling for analytics, support triage, claims, Stripe webhooks, outreach, cache management, and system logs.

Technical highlights

• Buyer matching pipeline using outbox events, SQL prefilters, FOR UPDATE SKIP LOCKED claiming, and a shared scoring evaluator.
• Real-time deal, offer, notification, inquiry, and team chat updates through SignalR with Redis-backed scale-out support.
• Document templates, generated PDFs, signatures, and production signing support through Azure Key Vault-backed certificate infrastructure.
• SSR-first public pages, Core Web Vitals baselines, manual Vite chunking, dynamic imports, self-hosted fonts, image proxy resizing, and optional Blob Storage rendition caching.
• Structured logging with Serilog, Datadog forwarding, OpenTelemetry support, persisted monitoring logs, frontend diagnostics, health checks, and support snapshots.

The strongest technical story is the combination of product breadth and operational depth: Equimeta is not just a listing site, but a maintainable production system designed for real users, complex transactions, and long-term platform growth.

Daedalus is a HubSpot portal audit and configuration platform for teams whose CRM has become expensive, cluttered, and hard to trust.

The product turns consultant-style portal cleanup into a guided software flow: connect the portal, review the recommended configuration, approve or modify each change, and deploy with rollback available.

The experience is focused on business outcomes: pipelines that match how sales actually works, properties that have a reason to exist, workflows that move lifecycle stages forward, dashboards the board can trust, and a CRM the team will use.

The Cyber AB site supports the organization responsible for coordinating key parts of the U.S. cybersecurity and CMMC ecosystem.

I oversaw architecture, planning, development, and maintenance for a custom e-commerce and integration platform with numerous third-party services. The work included reverse engineering the previous production database, translating data into the new system, and coordinating production integrations across the stack.

The result was a more connected platform for commerce, credentialing workflows, customer operations, training, CRM processes, background jobs, and reporting.