Project evidence for reviewers

Case Studies & Project Evidence

A hiring-manager friendly index of the systems, research work, deployment choices, costs, and tradeoffs behind the projects I want reviewers to inspect.

Read PromptTriage Browse all studies

Systems Working software, demos, repositories, and source-backed implementation notes.

Research Benchmark runs, evaluation summaries, charts, cost data, and limitations.

Operations Deployment history, logs, incident lessons, health checks, and rollback plans.

Hiring signal Fast summaries that lead to deeper proof without making reviewers hunt.

Case study library

Separate pages, separate evidence trails

PromptTriage has the complete structured page now. The other projects are listed as dedicated units so they can become full template-backed studies without changing the navigation model later.

Complete case study

PromptTriage

RAG-powered prompt engineering platform and Azure ML research study

Full case study with current architecture, Supabase Auth, Stripe billing, Pinecone retrieval, Azure ML benchmark evidence, release diagrams, cost notes, failure modes, and remaining evidence gaps.

28K+ prompt corpus 8 release diagrams EUR 178.12 Azure VM spend 38.19h longest benchmark

Open case study Repository

Source verified

Kaelux.dev

AI engineering agency platform

Service site and technical platform for AI engineering work, including a diagnostic agent, project hub, wiki strategy, SEO/GEO content, and deployment documentation.

Live platformRepositoryDiagnostic UXDeployment notes

Live site Repository

Part one published

ViperMesh

Deterministic Blender agent harness and MCP benchmark study

Benchmark comparing the ViperMesh Blender harness against the Anthropic x Blender MCP server baseline, including elapsed time, estimated token allocation, deterministic tool use, neutral LLM visual scoring, and EvoRig harness-evolution evidence.

6 of 7 speed wins2.534x mean speedup+8.19 visual points-90.91% reduced token usage

Open case study Repository

Source verified

n8n Automation Atlas

Workflow library and automation evidence base

Curated automation library with ready-to-import n8n JSON exports, license preservation, and an explorer path for builders who need practical workflow examples.

Workflow curationImport UXJSON corpusLicense notes

Atlas repository Workflow corpus

Documentation standard

Next pages will follow the same case-study template

Each future page should cover executive summary, problem, constraints, requirements, architecture, security, deployment, operations, cost, results, tradeoffs, failure modes, improvements, repo/demo links, interview explanation, and resume bullets.

View portfolio GitHub