Production AI.
Not Prototypes.

Two years of building, deploying, and operating AI systems. Every project here is live, in production, and has been through the kind of iteration that only comes from real users and real failures.

View CV → See BiasFeed Live
BiasFeed homepage with latest bias analyses
BiasFeed bias analysis with highlighted framing differences
BiasFeed admin panel with content management
BiasFeed n8n workflow pipeline
01 Live at biasfeed.com

BiasFeed: AI Media Bias Analysis

An automated pipeline that deconstructs news articles into neutral facts, then rebuilds them through three AI personas (progressive, conservative, neutral) to show how the same information gets framed differently across the political spectrum. Full admin panel with content management, keyword tracking, analytics, and a direct pipeline trigger. Runs autonomously with daily output.

How It Works

  • Source Ingestion: News articles ingested automatically from selected sources
  • Fact Extraction: AI strips opinion and charged language, distilling verifiable facts
  • Multi-Persona Generation: Three separate AI personas rebuild the story from the same facts
  • Bias Annotation: Interactive "Show Bias" toggles reveal exactly where and how framing differs
  • Admin Console: Full content management, analytics, keyword configuration, and manual pipeline triggers

What It Demonstrates

  • Multi-model orchestration: different AI models for different pipeline stages based on capability, cost, and reliability
  • Production workflow design: error handling, scheduling, monitoring, autonomous operation
  • Responsible AI: using AI to increase transparency rather than obscure it
n8n Multi-model AI Automated pipelines PHP Admin panel
Chester's blog with AI-written posts
Social Content Hub with LinkedIn publishing pipeline
Uplink research briefing engine
Relay device-to-device file sharing
Okavyx status dashboard with infrastructure monitoring
02 Live at okavyx.ai

Okavyx: Agentic AI Platform

A full agentic AI platform built from scratch. 14 production services including text and voice chat, sub-agent orchestration, infrastructure monitoring with self-healing, automated content pipelines, and a blog written by the AI agent. Persistent memory across sessions, multi-model routing, and security built into the architecture from day one.

Architecture

  • Persistent Memory: Long-term context across sessions, semantic search, temporal decay
  • Multi-Model Orchestration: Routes tasks to different AI models based on capability, cost, and complexity
  • Sub-Agent Delegation: Complex tasks broken into subtasks delegated to specialised coding agents
  • Real-Time Voice: Sub-second voice interaction via ElevenLabs, multiple interface options
  • Infrastructure Monitoring: Custom status dashboard, 50+ health collectors, Telegram alerting, self-healing
  • SSO & RBAC: Google OAuth, JWT cookies, role-based access across all 14 services
  • Autonomous Operations: Cron-driven tasks, scheduled workflows, background operations
Agentic AI Multi-model Voice interface Docker Node.js PostgreSQL Redis
Status dashboard with monitoring panels
Status dashboard infrastructure monitoring detail
03 Live at status.okavyx.ai

AI-Integrated Infrastructure Monitoring

A custom monitoring dashboard with 50+ health collectors across services, containers, SSL, DNS, and infrastructure. AI-driven alert analysis distinguishes expected changes from genuine anomalies. Self-healing for known failure patterns, Telegram alerting, and real-time visibility into the entire platform.

Capabilities

  • Automated Health Checks: Scheduled probes across services, containers, SSL, DNS, and infrastructure
  • AI-Driven Alerts: Contextual reasoning about whether changes are expected, correlated, or anomalous
  • Self-Healing: Automated recovery for known failure patterns before escalating to a human
  • Multi-Panel Dashboard: System resources, reachability, web security, git repos, memory, containers, and more
Infrastructure monitoring AI alerting Self-healing Docker Node.js
YouTube Shorts dashboard with category selection
YouTube Shorts n8n pipeline workflow
04 Live on YouTube @okavyx

Automated YouTube Shorts Pipeline

A fully automated pipeline producing publish-ready YouTube Shorts from a single topic input at ~$0.03 per video. Script generation (GPT-4.1 Mini), voice synthesis (ElevenLabs), parallel video sourcing with AI-judged selection (Gemini Flash), ffmpeg assembly with Ken Burns effects and word-by-word subtitle highlighting, and a human-in-the-loop approval workflow.

Pipeline Stages

  • Script Generation: GPT-4.1 Mini produces structured scripts with narration, visual queries, SEO metadata, and hashtags
  • Voice Synthesis: ElevenLabs TTS with character-level timestamps for word-by-word subtitle sync
  • Video Sourcing: "Video Department" sub-workflow searches Pexels, Pixabay, and Openverse in parallel; AI judge picks the best clip per segment
  • Assembly: ffmpeg handles portrait cropping, crossfades, Ken Burns on stills, subtitle highlighting, background music mixing
  • Upload & Approval: Private YouTube upload, Pushover notification, one-tap approve/reject with SSE live feedback
  • 7 Content Categories: Did You Know, Myth vs Reality, How It Works, History in 60s, ELI5, Hot Take, Top 3
n8n ElevenLabs OpenRouter YouTube API ffmpeg $0.03/video
05

AI Workflow Orchestration & Automation

Cross-platform AI workflow orchestration: automated LinkedIn content with Telegram approval workflow, AI image selection service (parallel stock API queries + AI judge), cron-driven content generation, security-hardened API endpoints, and watchdog processes ensuring reliability without supervision.

Systems

  • Social Content Pipeline: Cron-scheduled LinkedIn drafting with inline approve/reject/edit buttons via Telegram
  • AI Image Battle Royale: Reusable workflow querying multiple stock APIs in parallel, AI selects the best match
  • Multi-Step Task Chains: Conditional logic, error handling, rollback capabilities
  • Cross-Platform Integration: Messaging, monitoring, version control, and infrastructure management
n8n Workflow automation LinkedIn API Cron scheduling

Microsoft 365 & Azure AI

Enterprise solutions built entirely within the Microsoft ecosystem. Azure Functions, Graph API, SharePoint, Entra ID. All running on free or near-free tiers.

AskSite knowledge bot answering questions from SharePoint documents
AskSite bot in Microsoft Teams with document citations
07 Live in production

AskSite: Copilot Alternative at Near-Zero Cost

A Microsoft Teams bot delivering the same core experience as $30/user/month Copilot: ask a question in natural language, get an AI-generated answer citing your SharePoint documents. Built on Azure free tiers with pay-per-token OpenAI. For a 20-person company, this replaces $7,200/year in licensing.

Architecture

  • Teams Integration: Direct chat or @mention in any channel. No new tools to learn.
  • Azure AI Search: SharePoint documents indexed and chunked for semantic retrieval
  • Azure OpenAI (GPT-4o): RAG pipeline with document-grounded answers and source citations
  • Zero Infrastructure Cost: Azure Bot Service (free), Azure Functions (consumption), AI Search (free tier)
Azure OpenAI Azure AI Search Bot Framework Microsoft Teams SharePoint Entra ID
Tenant Health Check traffic-light report
08 Live, running daily

M365 Tenant Health Check

Automated daily security auditing for Microsoft 365 tenants. Runs inside the client's own Azure tenant, checks MFA coverage, stale accounts, admin sprawl, email authentication, and password policies. Emails a traffic-light HTML report designed for non-technical business owners. Built as a deployable product for MSPs and IT consultants.

Azure Automation Microsoft Graph API Entra ID Security auditing Python
09 Live, client deployment

Automated Document Processing Pipeline

Azure Function monitoring a SharePoint folder for unstructured diagnostic files. Intelligent multi-line parsing extracts structured data from complex formats (nested sections, variable layouts), writes to SharePoint Lists via Graph API. Staff get instant search and filtering without changing their existing workflow. 108 out of 108 fault types validated.

Azure Functions Microsoft Graph API SharePoint Lists PowerShell Data pipeline
TimeSheet time entry form in Microsoft Teams
TimeSheet invoice generator with unbilled summary
Generated invoice with line items and email delivery
10 Live, daily use

Time Tracking & Invoicing on M365

End-to-end time tracking and invoicing built entirely on Microsoft 365. SharePoint Lists for capture (mobile, desktop, or Teams tab), webhook-driven hour calculation, one-click PDF invoice generation with Entra authentication, automatic SharePoint storage and email delivery via Graph API. Replaces SaaS subscriptions with infrastructure you already pay for.

SharePoint Azure Functions Microsoft Graph API Entra ID PDF generation Webhooks

These projects span two domains: independent AI systems and enterprise Microsoft 365 solutions. The common thread: production systems solving real problems, at minimal cost, using the right tool for the job.

A pattern runs through the AI projects: agentic AI directed to build its own automation infrastructure. The design decisions are human: which models, which APIs, what the cost target is, where human approval gates belong. The AI agent implements those designs, building workflows that handle high-volume production work at near-zero marginal cost.

Every project listed here is live, in use, and has been iterated based on real operational experience.

Get in Touch

Interested in what you see? Whether it's AI strategy, automation, or just a conversation about what's possible.