Projects

Every engagement is a reference for what we can deliver. Here is what clients have asked for and what shipped.

5 Projects

Area: All · Data Intelligence (3) · AI and Intelligent Automation (5) · Digital Experience Engineering (1) · Infrastructure and Platform Engineering (1) · Interactive and Gaming Solutions (2) · Embedded and IoT (1)

AI Orchestration Platform

A client ran one multi-step technical workflow: fetch data from APIs, analyse it, generate code, deploy, and report. Every step used a different tool with a human handoff between them, so one run cost four hours of attention.

  • Headline: Autonomous Workflows Without a Middleman
  • Outcome: A four-hour manual process now takes 15 minutes of review time.
  • Stack: Python, Hermes Agent, MCP, Docker

Area: AI and Intelligent Automation

Character AI Bot

A character AI bot with a strict hexagonal architecture, agentic memory, swappable personality modules, and conversation state that survives a restart.

  • Headline: Character AI Bot
  • Outcome: Personalities deploy, swap, and extend independently of the core system.
  • Stack: Python, FastAPI, Redis, Hexagonal Architecture

Area: AI and Intelligent Automation

Local RAG Pipeline

A retrieval-augmented generation pipeline that runs entirely on the client's own hardware: document ingestion, local embeddings, vector search, and LLM answers with no external API call.

  • Headline: Local RAG Pipeline
  • Outcome: Documents are queried in plain language without leaving the premises.
  • Stack: Python, Vector Database, Docker, Local LLM

Area: AI and Intelligent Automation

Open WebUI Theme Engine

A theming system for Open WebUI that generates CSS from JSON definitions, with per-component styling, gradients, glows, and no limit on how many themes a site can carry.

  • Headline: Open WebUI Theme Engine
  • Outcome: Organisations deploy a branded AI chat interface that matches their own design system.
  • Stack: CSS, JavaScript, JSON, Open WebUI

Area: AI and Intelligent Automation

Speech-to-Text Pipeline

Meeting transcription with speaker identification, on technical vocabulary that off-the-shelf services get wrong.

  • Headline: Meeting Transcription That Knows Who Said What
  • Outcome: Past meetings are searchable by speaker and by topic.
  • Stack: Python, NVIDIA NeMo, ASR

Area: AI and Intelligent Automation