Project proof
Production work across chat, workflow products, commerce, AI agents, documents, and search.
These are the kinds of systems I like building: connected to real users, shaped by backend constraints, and measured by operational outcomes.
ChatsX
20K WebSocket connection workload, Meta webhook normalization, unified inbox state, and LLM-assisted agents across Facebook, Instagram, and WhatsApp.
20K WS - Channels - Meta Graph
TradeX
Multi-tenant trading platform work across binary flows, wallets, ledgers, liquidity, orders, market data, bots, and admin controls.
Ledger - Liquidity - Bots
DroidFlow
Automation engine with retryable workflows, explicit API boundaries, operator state, recovery paths, and client-facing control.
Retries - APIs - Recovery
04Erixa
Per-client WhatsApp commerce bot with Shopify order sync, screenshot fraud checks, OCR, ELA, TruFor, CAT-Net, and Claude Vision review.
Forensics - OCR - Shopify
05LadybugWorth
Slack-monitoring AI agent that extracts decisions, builds EOD reports, summarizes team threads, and flags bug-risk signals before release.
Slack - Agents - Risk Signals
More shipped work
Proof beyond the headline projects.
ResumeFolio
Replaced a paid .NET document vendor with a Python-native DOCX-to-HTML engine, custom CSS injection, OCR-assisted reconstruction, and pixel-accurate resume output.
PixelioWall
Image gallery backend managing 60K+ assets with semantic vector search, pgvector indexing, PostgreSQL query tuning, and S3 storage paths.
Multi-tenant SaaS
Schema-isolated tenant architecture, Django APIs, React admin surfaces, Docker deployment, AWS EC2 operations, and zero cross-tenant data incidents.
Production recovery work
Saved 2 startups by reverse-engineering provider WebSocket protocols into direct Python clients; helped another avoid a critical Django RCE exposure.
Need this kind of build?
Send the system, stack, bottleneck, and deadline. I will reply with the first technical move I would make.