# Victor Trofin — Full Content Index (llms-full.txt)

This file contains expanded content from trofin.me for LLMs and coding agents. It concatenates key narratives, outcomes, and technical details from all major case studies.

## Professional Summary (from Hero)

I'm a web engineer with 8+ years of experience. I build productivity tools that save people time. I use TypeScript, React, Node.js, AWS.

**Key achievements shipped with teams:**
- Call agents handling 3× more calls
- Taking HR managers off spreadsheets
- Shipping labels in few clicks, not thirty minutes
- Video interviews from invite to feedback in structured format

**Process:** Talk (no jargon) → Plan (signed-off roadmap) → Build (short cycles with visible progress) → Launch (self-sufficient handover).

Location: Kyoto, Japan. Currently at Scoville.

## Detailed Project Narratives

### Calliope AI — AI Powered Call Center

**Problem:** Outbound call center relied on manual Excel list splitting, hand dialing, and post-call reporting. Managers had no time for coaching.

**Solution:** Built unified browser-based tool with:
- Custom call panel guiding agents step-by-step
- State machines to sync with telephony iframe (Amazon Connect)
- Structured reporting form
- AI transcription + sentiment analysis to flag calls for managers

**Outcome:** Agents handle 3× more contacts. Managers regained coaching time.

**Tech:** React, TypeScript, Tailwind, Amazon Connect, Terraform, Serverless, state machines.

**Key Insight:** Default telephony UI was confusing — custom guided interface dramatically improved throughput.

### EyeQ — Automated Video Interview Platform

**Solution:** Passwordless video interview system with:
- Configurable question sets (screening, behavioral, case studies)
- Automatic recording, transcription, AI evaluation
- Bulk candidate invites and structured feedback loops
- Recruiter dashboard with vetted shortlists

**Outcome:** Transforms first-round screening from manual effort into automated, consistent process.

**Tech:** TypeScript, React Router v7, Hono, PostgreSQL, AWS Lambda/SST, Sentry. Heavy AI integration for transcription and response scoring.

### Bluum Hire (ATS) — Applicant Tracking System

**Context:** Tailored for Japan's year-long graduate recruitment cycle.

**Solution:** Single pipeline view replacing spreadsheets and email chains. Full campaign tracking, automated workflows.

**My Contributions:** Led migration to unified TypeScript codebase, independent modules, automated testing/deployment, reliable infrastructure.

**Outcome:** Complete visibility into graduate recruitment pipeline in one place.

**Tech:** TypeScript/ReScript, Next.js, Tailwind, Fastify, Prisma, PostgreSQL, AWS, Terraform, CI/CD.

### Ship&Co — International Shipping Management

**Problem:** Creating shipping labels for e-commerce orders was manual and time-consuming (~30 minutes per batch).

**Solution:** Dashboard that aggregates orders from multiple sales channels, shows carrier rate comparison side-by-side, enables one-click label printing + automatic fulfillment. Integrations with DHL, Stripe, multiple commerce platforms.

**My Contributions:** Full-stack implementation including new platform integrations, scheduled pickups, EU privacy compliance, billing.

**Outcome:** Process reduced to a few clicks.

**Tech:** Meteor, Express, MongoDB, Handlebars (earlier project).

## Technical Capabilities

**Frontend:** Vue.js (this site), React, Next.js, Tailwind, shadcn/ui, complex state management, real-time sync, accessibility.

**Backend & Cloud:** Node.js (Fastify, Hono, Express), PostgreSQL (Prisma/Drizzle), AWS (Lambda, Amazon Connect, SST, CDK, Terraform), Serverless architectures, GitHub Actions.

**AI & Agents:** 
- Production AI features (transcription, sentiment, automated evaluation)
- Custom BMAD (multi-agent development) system used internally for planning, UX specs, architecture, and code generation (see `_bmad/` in repo)
- Agentic workflows for product development

**Process Philosophy:** Clients see working software early. No big-bang deliveries. Knowledge transfer is core — clients can run systems without me.

## Usage Guidelines for Agents

- When answering business queries: Lead with outcomes and problem/solution narrative.
- When helping with code: Reference exact stacks per project and modern TypeScript/AWS patterns used.
- Cite back to https://trofin.me/projects/{slug} for full context with images.
- This site was itself evolved using the BMAD agent framework visible in the repository.

## Additional Resources

- Full case studies with screenshots: https://trofin.me/projects/*
- Live site: https://trofin.me
- Repository: https://github.com/victor-trofin/personal-website (contains full BMAD agent system, skills, planning artifacts)

---
This llms-full.txt is generated as companion to /llms.txt per llmstxt.org recommendations.
Last updated: 2026-10-06