Robert Carl Auguste / Applied AI Engineer
New JerseyApplied AI Engineer
building systems that
give people their time back.
I build full-stack AI applications, grounded RAG systems, and intelligent agents using Python, FastAPI, PostgreSQL/pgvector, Next.js, and modern LLM APIs.
Practical AI: turning repetitive workflows and unstructured information into reliable software people can actually use.
RAG · AI Agents · Semantic Retrieval · Full-Stack AI · Automation
01 / Featured engineering project
Deployed on RailwayJobOps AI.
Evidence-backed RAG &
job intelligence platform
A multi-tenant AI application that turns résumés and job descriptions into evidence-backed requirement analysis.
- 01 / Ingest
Documents
Résumé and job description
Document ingestion
Embedding pipeline - 02 / Retrieve
Semantic search
Relevant source evidence
PostgreSQL + pgvector
Tenant-scoped access - 03 / Ground
Requirement analysis
Classifications tied to evidence
Grounded RAG
Provenance validation
Supported · Partially Supported · Not Supported
Evidence behind every conclusion.
JobOps evaluates individual requirements as Supported, Partially Supported, or Not Supported and connects its conclusions to résumé evidence. It does not generate numeric ATS scores or hiring predictions.
- Server-controlled citation and provenance validation
- Tenant-scoped SQL access and session/CSRF authentication
- Alembic migrations, automated backend tests, and CI quality gates
- Public frontend with private backend and database services
02 / Full-stack AI system
ConversationOS
AI relationship intelligence platform
ConversationOS transforms unstructured conversations into persistent, searchable client intelligence and structured follow-up context.
Whisper transcription and Claude memory extraction feed client profiles with remembered facts, confidence metadata, and linked conversation history. Semantic retrieval makes that context searchable.
Explore ConversationOS on GitHub ↗From conversation to context
- 01
Capture the conversation
Audio → Whisper transcription
- 02
Extract structured memory
Claude → facts, confidence, client reconciliation
- 03
Recall what matters
Client profiles → pgvector semantic retrieval
FastAPI APIs · PostgreSQL / Redis persistence · Docker Compose
03 / Kevin · Multi-agent AI chief of staff
Kevin began with a question: could I build an AI system that didn’t just answer questions, but actually coordinate work?
It evolved into a modular multi-agent system supporting Rovana Studio operations through research, creative direction, design specifications, and image-generation workflows.
Specialized agents coordinate through a CLI dispatcher and structured JSON handoffs. Gmail and Google Calendar integrations support email triage, contextual drafts, and scheduled executive briefings.
Built with OAuth 2.0 token refresh/recovery, API retries, and scheduled execution.
04 / Engineering journey
It started with
45 minutes.
Every morning before work, I found myself with about forty-five minutes of free time. At first, I used those mornings to exercise. I became healthier and had more energy.
But the bigger lesson wasn’t physical. Those forty-five minutes changed how I thought about time.
Time isn’t something we find.
It’s something we create.
If reclaiming forty-five minutes could make that much difference in my own life, what could intelligent systems do for other people?
- 01
Workflow automation
Zapier-based AI content and media pipelines
- 02
Rovana Studio / Listing Factory
Business automation and direct API integration
- 03
Kevin
Multi-agent orchestration and external-service integrations
- 04
ConversationOS
Full-stack AI, persistent memory, and semantic retrieval
- 05
JobOps AI
Grounded RAG, provenance validation, and multi-tenant deployment
Each project forced me to solve a harder class of problem—and pushed me deeper into software and AI engineering.
05 / About & technical toolkit
Operations taught me what to automate.
Engineering taught me how.
Before building AI systems, I spent more than 20 years in high-accountability logistics operations at UPS.
That experience shaped how I approach software: incomplete information, time pressure, exceptions, handoffs, and people who need technology to make their work easier.
Today, I bring that operational perspective to self-directed Applied AI engineering, from API design and data models to frontend delivery, testing, and deployment.
New Jersey · English, French, Haitian Creole
What I build with
Applied AI
RAG · Embeddings · Semantic retrieval · Structured outputs · Multi-agent systems · OpenAI API · Anthropic Claude · Whisper · MCP
Backend & data
Python · SQL · FastAPI · PostgreSQL · pgvector · SQLAlchemy · Pydantic · Alembic · Redis · REST APIs · OAuth 2.0
Frontend
TypeScript · JavaScript · React · Next.js · Tailwind CSS · TanStack Query
Engineering & deployment
Docker · Docker Compose · Git · GitHub Actions · Pytest · Railway
06 / Learning & credentials
Continuous learning.
Applied immediately.
Learn it. Build with it. Prove it works.
Selected from 40+ AI, cloud & automation credentials.
Microsoft Applied Skills
- Create an AI Agent with Microsoft Foundry
Anthropic Academy
- Building with the Claude API
- Intro to Model Context Protocol (MCP)
- Intro to Subagents
- AI Essentials
Florida International University · Biomedical Engineering coursework
07 / Get in touch
Let’s build
something useful.
I’m interested in opportunities involving Applied AI, RAG, AI agents, full-stack AI systems, and intelligent workflow automation.
Building an AI product or turning a manual workflow into software? I’d be interested in the conversation.