Open to AI Systems Engineering / Agentic Orchestration / AI Enablement roles

Scott Certain

AI Systems Engineer

AI Systems Engineer who builds agentic systems, proves they work with independent evaluation, and drives adoption.

Applaud TechnologiesGenesis10 / Dev10 Talent DevelopmentSpeedwayThe Coffing CorporationTotal Quality Logistics (TQL)United States Military

Or paste a job description →

scroll

Experience

Each role includes queryable AI context — the real story behind the bullet points.

GhostRyter · PickedUp · Buckeye Property Group

AI Systems Engineer (Independent)

Jan 2026 – Present

→Multi-agent orchestration. Designed and built GhostRyter, a multi-agent production platform in Claude Code: 17 agents across 4 pipelines governed by a single 523-line production specification, with per-agent blocking acceptance criteria verified programmatically before work advances to the next stage.
→Evaluation architecture. Built an architecturally firewalled evaluation agent that grades every draft on five dimensions (voice fidelity, strategic alignment, reader service, factual integrity, cross-format consistency) with no access to the production spec or generation agent configs — an LLM-as-judge design that prevents the evaluator from grading toward the generator's instructions.
→Reliability and operations. Queue layer for unattended batch processing with human checkpoints, git commit per pipeline run with non-blocking failure logging, SQLite research-references database with URL health monitoring, LanceDB vector store for semantic search across prior output, and DevContainer plus Docker isolation. A full run produces ~70 graded drafts across 6 formats in ~30 minutes of unattended runtime.
→Voice agents. Built three production-grade Retell AI voice agent designs across two LLM providers (Claude Sonnet 4.6, gpt-4.1-mini) for real estate front-desk reception and after-hours emergency maintenance triage, including a 39-node conversation flow with 9 explicit Fair Housing Act guardrail nodes and tool-calling for triage and lead-capture actions. Twilio telephony, n8n on Railway for post-call orchestration.
→Companion web app (pickedup.live). Next.js 14 / TypeScript site with browser-based real-time voice calls, Kit lead capture, libsql/Turso persistence, Netlify deployment, and HMAC-SHA256 webhook signature verification with 5-minute timestamp window enforcement.

Applaud Technologies, LLC

AI Software Developer / Curriculum Developer

Oct 2023 – Jan 2026

→Authored a Generative AI for Software Developers curriculum across two engagements for a nationally known coding bootcamp: first as Part 2 of an inherited co-authored course (5 modules), then as a full 12-module rewrite the same client commissioned after preferring this work's projects and technical coverage. Coverage: prompt engineering, RAG, LangChain/LangGraph multi-agent systems, LoRA fine-tuning of Falcon 7B and Gemma 3 4B with Ollama deployment, and LLMOps. Built and verified working reference systems for every coding activity, including an HR-assistant capstone (fine-tuned LLM, RAG, LangGraph orchestration).
→Designed, taught, and recovered a mis-scoped 8-week data engineering bootcamp (SQL Server through Azure Databricks ETL). When the client's assessment of student capability proved inaccurate in week 1, restructured the course on the fly using AI-augmented authorship against the Claude API, producing lesson content 1-2 days ahead of live delivery for eight straight weeks. Four of eight students placed and were performing well at the 3-month follow-up.
→Architected an LLM-integrated web application enabling non-technical users to generate course content with AI, and built reusable prompt systems in Anthropic Workbench for repeatable content generation.
→Modernized a legacy application to .NET Core Web API, Entity Framework Core, Angular, and NgRx for a large Ohio-based food service client.

Genesis10 / Dev10 Talent Development

Technical Instructor & Curriculum Mentor

May 2021 – Jul 2023

→Trained and mentored 100+ adult learners in Java, SQL, and React across five 12-week cohorts, delivering live instruction, side-by-side technical guidance, and code-level feedback. Graduates placed at major employers.
→Conducted 700+ JavaScript code reviews for the recruiting team, holding a consistent technical quality bar across cohorts — the same validation discipline now applied to reviewing AI-generated code for correctness, security, and maintainability.
→Trained 20 Data Engineer Associates in advanced SQL: temporary tables, common table expressions, views, and stored procedures.

Speedway, LLC

Programmer Analyst / Application Support Analyst

Apr 2018 – May 2021

→Administered Azure DevOps and Octopus Deploy for the team's in-house applications, owning build and release pipelines across development, testing, and production.
→Migrated 14 .NET/SQL applications from Marathon Petroleum's network to Speedway's infrastructure, conducted code reviews for 3 contract developers, and deployed 10 custom ASP.NET/SQL applications supporting a supply chain management system upgrade.
→Served as Incident Manager for 10 major system outages, coordinating troubleshooting teams and keeping leadership informed in plain terms throughout each incident.

The Coffing Corporation

Software Developer

Nov 2017 – Apr 2018

→Maintained the WSDL interface for Oracle Primavera connecting US Air Force IT Systems Portfolio Management to DoD systems.

Total Quality Logistics (TQL)

Software Developer

May 2016 – Nov 2017

→Developed and maintained a C# / SQL Server logistics application used by 4,000 brokers managing 1,000,000+ freight loads annually. Implemented a real-time market rate search feature enabling accurate quotes, and eliminated 75% of existing technical debt during a holiday code freeze.

What I'm good at — and what I'm not

Skills, ranked honestly. The gaps column is the one most candidates won't show you.

✓

Strong

Deep experience. Hire me for these.

Multi-agent orchestration
Agent evaluation architecture (LLM-as-judge)
Claude Code (agents, sub-agents, skills, slash commands)
Production validation of AI-generated code
System specification writing
Prompt engineering / Anthropic Workbench
Agent guardrail and constraint design
Voice AI agent design (Retell AI)
Honest scope calibration
LangChain / LangGraph multi-agent orchestration
RAG systems
Multi-provider LLM selection
Anthropic API (Claude Opus / Sonnet)
LoRA fine-tuning (small open-source models)
Technical writing and documentation at scale
AI enablement and developer training
AI-augmented production at scale
Quantified output specification

○

Moderate

I can contribute. Not the expert.

Webhook integration and HMAC-SHA256 signature verification
Queue-based batch processing
n8n workflow orchestration
Twilio / telephony orchestration
SQL Server
PostgreSQL / Supabase
SQLite / libsql / Turso
C# / .NET / ASP.NET Core
Entity Framework Core
Angular / NgRx
React / Next.js
Python
JavaScript / TypeScript
Java / Spring Boot
Docker / DevContainers
Git / GitHub
CI/CD pipeline administration
Azure DevOps
Octopus Deploy
Pandas
Azure / Azure Databricks
Incident management
Military leadership / mentorship

✗

Gaps

Don't hire me for these.

Senior IC classical programming
ML engineering / MLOps
Large-scale production engineering
Mobile development
Frontend design / UX
Consumer product / growth experimentation
Compliance practice (regulatory practitioner)

Honest Fit Assessment

Paste a job description. Get an honest assessment of whether I'm the right person — including when I'm not.

The AI uses my full private context to grade fit honestly. No prompts to please you.

Why this exists. This signals something completely different than "please consider my resume." You're qualifying them. Your time is valuable too.