LWC AI GROUP / CAPABILITIES CV
An independent practice connecting agent systems, software, physical interfaces and creative production.
lwcai.group · lwcai87@gmail.com · github.com/lwc2637
CAPABILITIES
Task-specific tools, execution harnesses, local model integration and reviewed personal context.
Document extraction, visual inspection, speech recognition and structured model output.
Responsive web interfaces, application APIs, durable jobs and complete user journeys.
Embedded voice controls, local audio bridges and Linux computer environments.
Character references, AI shot direction, editing, image treatment and website motion.
Restricted tools, input validation, source traceability, recovery and human review.
SELECTED WORK
An agentic market watcher and paper-trading bot, connecting market observation with automated decisions and simulated trades.
Contribution: Product concept and agentic workflow design.
Agentic automation · Market monitoring · Paper trading
A document review prototype that brings PDF extraction, bounded vision inspection and human decisions into one traceable workflow.
Contribution: Application design, AI integration and workflow engineering.
Python · FastAPI · React · TypeScript · PostgreSQL · Hermes · LM Studio · Docker
A physical Linux host with desktop, browser and camera tools, backed by a command harness, durable jobs and recovery tooling.
Contribution: Host provisioning, harness engineering and agent integration.
Raspberry Pi 5 · Ubuntu · Python · systemd · Hermes · Docker · Xfce
An embedded voice companion that connects a small physical device to local speech recognition, a language model and spoken replies.
Contribution: Embedded interface, voice pipeline and local application integration.
M5Stack StickS3 · C++ · PlatformIO · Python · Faster Whisper · LM Studio · Windows SAPI
A conversational application exploring personal context, approved memories and the foundations of a more individual AI assistant.
Contribution: Conversational product development and model integration.
Next.js · TypeScript · OpenRouter
Website and app work spanning a multilingual taxi-interface demo and an image-led hospitality website with a cinematic media treatment.
Contribution: Experience design, frontend development and visual production.
HTML · CSS · JavaScript · TypeScript · Astro · GSAP · Higgsfield · FFmpeg
A 30-second film experiment combining reference-led characters, action staging, timed cuts and a final title frame.
Contribution: Creative direction, character references and editorial timing.
AI video generation · Character references · Video editing · Audio
Five container-based computer environments with internal MCP tooling. Each Neuro-Agentic Base gives an LLM and its harness a workspace to build in; bases can connect to form Hive-Minds.
Contribution: Agent environment architecture, MCP tooling and harness design.
Containers · MCP · Agent harnesses · Computer environments · Hive-Mind coordination
WORKING APPROACH
Start with the user task. Connect model output to a clear workflow. Test meaningful behaviour, make failures visible and document the boundaries of what has been demonstrated.
Project status is part of the evidence. Prototypes and demonstrations are identified throughout this CV; detailed case studies describe current limitations and planned integrations.
This document summarises capabilities and selected projects. The online portfolio contains the current case studies and media.