Jordan HenningBook a call
AI engineering & delivery leadership

I build multi-agent AI — backed by a decade leading federal IT delivery.

An AI engineering and delivery leader who ships the systems himself. I design, build, and run multi-agent AI systems end-to-end: RFP Factory (in production at Quadratic) compresses 40-hour federal proposal cycles into 2 hours of human review; a multi-agent trading desk built end-to-end (personal R&D); and the multi-agent marketing system that runs my ShadowEdge Tools product line. That hands-on depth sits on 17 years of federal service at the Social Security Administration, nearly a decade of it leading IT delivery — as Branch Chief (2022–2025) I ran 12 direct reports (incl. 2 team leads) and set IT governance for a 340-technician operation across 170 nationwide offices at 99.9% availability; earlier, as a FAC-P/PM-IT–certified IT Project Manager (2016–2022), I led delivery of $200M+ in IT projects across matrixed teams, on-time and under-budget. Public Trust background investigation (held 2008–2025; eligible to reinstate). Combat veteran (101st Airborne). I build the AI myself, and I've led IT delivery at federal scale for nearly a decade — so I'm ready to lead the AI engineering team that ships it, with an authority most leaders can't bring.

The hard part of an AI product isn’t the model — it’s reading the workflow fast enough to know what to build.

Book 15 minutesDownload resumejordanhenning32@gmail.comGitHub
Public Trust · Reinstatement Eligible•Former FAC-P/PM-IT•SSA Commissioner Award•MBA, Malone
Jordan Henning, executive portrait
Today
Jordan Henning
AI Engineering & Delivery Leader · Builds multi-agent AI systems hands-on · a decade leading federal IT delivery at SSA · $200M+ delivered · Former FAC-P/PM-IT
Vision

If you can dream it, you can build it. I'm here to do both.

Technical work in 2026 demands visionaries more than ever — leaders who can hold the entire picture of a product, from first sketch to live deployment, and develop the whole arc with intent.

We're past the era when ideas were the bottleneck. The tools are here. The compute is here. The agents are here. What's scarce now is the operator who can see the end-state, build a path to it, and ship it without losing the original conviction.

I'm that operator — an AI delivery leader who ships the systems himself and has led delivery teams at federal scale. I read architecture, review code, and tune agent systems without a translator, and I've run multi-tier teams to a 99.9%-uptime bar. The next role is leading AI delivery as a hands-on player-coach — setting architecture direction, leading the engineering team, and owning the delivery outcome — grounded in a decade of real federal delivery discipline.

Read the full argument
Federal delivery record

The delivery record, in numbers.

Every number below is from real federal and commercial work — hover or tap any metric for the source.

340
Field IT operation governed
As Branch Chief: 12 direct reports (incl. 2 team leads) and IT governance, policy & standards authority over a 340-technician field operation across 170 SSA Hearings Offices, serving 7,000+ employees
170
Nationwide offices unified
End-to-end IT service delivery across all SSA Hearings Offices
99.9%
System availability
Sustained for mission-critical platforms serving millions of users
20%
Downtime reduction
Through proactive monitoring, dashboards, and ITIL-driven CSI
$200M+
IT projects delivered
Cumulative IT project value over a 5-year IT Project Manager tenure (2016–2022) at SSA — on-time, under-budget
Career arc

From one combat tour to AI delivery leadership.

A 25-year arc through the 101st Airborne, federal IT delivery at scale, and now building and leading AI delivery. Click any milestone for the detail.

    • Designed and ship RFP Factory — a role-separated, multi-agent proposal-automation pipeline (research → strategy → drafting → dual-model compliance review) that compresses 40-hour proposal cycles into 2 hours of human review (~20×), in production with grounded RAG and FAR/compliance guardrails.
    • Built the multi-agent marketing/CRM system that runs the firm's content and lead operations end-to-end behind human-approval safety gates.
    • Operate a FedRAMP-aligned delivery posture — AI builds run alongside ATO-aligned controls so the federal compliance baseline is the starting point, not an afterthought.
    • Embedded the firm as a subcontractor to Nava on CMS modernization; won two PA Invitation-to-Qualify (ITQ) vehicles and qualified into the PA Small Disadvantaged Business program.
    • Lead cross-sector growth strategy across federal and commercial markets — pricing, solution scoping, and capture.
    Multi-Agent AIAI DeliveryFedRAMPNava / CMSFederal + Commercial
Selected builds

Built these myself. This is the work I lead.

Every system here began as a workflow Jordan had already run — which is the point: what actually makes an AI product work is understanding the process deeply and quickly enough to know what deserves to be automated, where it breaks, and where a human stays in the loop. These are self-directed multi-agent systems he designed, coded, and operates hands-on (RFP Factory runs in production at Quadratic) — the depth behind the leadership. The next role is leading AI delivery as a player-coach: setting the architecture direction, leading the engineering team, and owning the delivery outcome — with the authority that comes from having built and shipped the work himself.

RFP Factory

In production · Quadratic Digital
Multi-agent proposal automation
Outcome
40h → 2h per proposal · 20× faster

An end-to-end system that ingests federal and commercial RFP documents and produces polished, SME-ready proposal drafts. Specialized agents handle research, strategy, drafting, and compliance review in parallel — compressing a 40-hour proposal task into 2 labor hours of human review, a 20× reduction in cycle time, with a human always in the loop on the final pass.

  • Decomposes unstructured RFP packets into structured response workstreams and compliance matrices
  • Orchestrates research, strategy, drafting, and review agents concurrently
  • Routes each task to the LLM provider that performs best for it (Anthropic, OpenAI, or Gemini)
  • Persists project state in a relational store with full audit history for SME review
Architecture
Human-checkpointed pipeline · grounded RAG · FAR-compliant · multi-provider
  1. RFP UploadPDF · DOCX
  2. Intake
    compliance_matrix · validator · metadata
    SonnetHaiku
  3. StrategyShortfall + gap analysis
    shortfall_strategist
    Opus
  4. Outline + Team
    outline_agent · team_composer
    Sonnet
  5. CostDual-pipeline market research
    market_researcher · cost_analyst · cost_reviewer
    GeminiClaudeGPT-5.5
  6. DraftPer-section with auto-resolve
    writer_team · needs_human_resolver
    SonnetHaiku
  7. Review (auto-loop, cap 4)
    reviewer_a · reviewer_b
    OpusGemini
  8. PolishCross-section consistency
    final_polish_detector · final_polish_applier
    Gemini ProSonnet
  9. Export DOCXcompile_proposal_to_docx
Python 3.12+Anthropic + OpenAI + GeminiSQLAlchemy 2.0AlembicNiceGUIThreadPoolExecutorPrompt engineering

Multi-Agent Dev System

Self-directed · builds the systems on this page
An entire software-engineering org, modeled as agents
Outcome
The engineered process behind every other system here

A pipeline that models a whole dev organization as role-separated AI agents — architect, coder, adversarial tester, bug-fixer, a security reviewer, and a senior QA sign-off — with deterministic verification before any code ships. It encodes how a real engineering team ships safe, quality software: separation of duties, independent review, security and QA gates, and proof-before-done. This is the opposite of 'vibe coding' — it's the engineered process that makes AI-built code trustworthy, and it built the other systems on this page.

  • Role-separated agents with independent adversarial review — no agent grades its own work
  • A security gate (BLOCK/CLEAN) and a senior QA sign-off (APPROVE/BLOCK) before anything ships
  • A deterministic verifier runs the real tests/builds/lint — 'looks good' from an LLM is never enough
  • Multi-provider routing (Claude plans and reviews, GPT-5.5 codes) with institutional-memory lessons carried across runs
Architecture
Plan → code → adversarial test → security gate → QA sign-off → deterministic verify
  1. Requirementshuman, once
  2. Architecttraceable spec + plan
    architect
    Claude
  3. Coderimplements every TODO
    coder
    GPT-5.5
  4. Tester ↔ Bug Fixeradversarial review + repair loop
    tester · bug_fixer
    ClaudeGPT-5.5
  5. Security Reviewtrading · fintech · secrets gate
    security_reviewer
    Claude
  6. QA Sign-offfinal shippability gate
    qa_reviewer
    Claude
  7. Weakness Routerroutes gaps → smallest reliable fix (after every stage)
    weakness_router
    Sonnet
  8. Deterministic Verifyruns real tests / builds / lint — proof before done
PythonClaude (Opus / Sonnet)GPT-5.5 / Codex CLIMulti-agent orchestrationDeterministic verificationMCP

Futures Bot

Personal R&D · multi-agent trading desk
Autonomous multi-agent trading desk
Outcome
Solo-built, end-to-end multi-agent system

A multi-agent trading desk that ingests futures market data, generates signals, manages risk, and manages orders autonomously — built end-to-end as a personal R&D platform for pressure-testing multi-agent design patterns under live-market conditions. The signal here is the engineering: specialized agents, hard-coded risk guardrails, and operator-level observability, designed and built solo.

  • Real-time futures market data ingestion and feature engineering
  • Specialized agents for signal generation, risk management, and order execution
  • Programmatic order placement with hard-coded risk guardrails
  • Runs autonomously with operator-level observability
Architecture
Market-regime engine · risk-gated order protocol · operator observability
  1. Market DataMES · MCL · NinjaTrader feed
  2. Regime Enginev3 classifier · transition tracking
    Deterministic
  3. Strategy Scanners10+ deterministic pattern detectors
    Breakout · Retest · ORB · RSI(2) · Liquidity Sweep · BB-RSI · Cup-Handle · BOS Order Block · AVWAP
    Deterministic
  4. Signal Rankingzone_guard · ranking
    Python
  5. 5-Agent CommitteeTwo-stage blind debate + synthesis
    Technical Analyst · Bull · Bear · Risk Manager · Devil's Advocate
    OpenAIClaudeGemini
  6. Risk GatesL2 walls · news adjacency · session · 6-tick scratch
    Python
  7. Order ExecutionProgrammatic SIM/live placement · scaled exits
  8. Causal Post-MortemEvery loss → root cause → regime params
    batch_post_mortem · causal classifier
    Haiku
PythonMulti-agent orchestrationReal-time market dataAlgorithmic executionRisk management

Résumé Engine

Self-directed R&D
Grounded multi-agent resume & cover-letter system
Outcome
Every claim traced to a verified fact before export

A multi-agent system that writes job-tailored resumes and cover letters grounded in a knowledge base of real experience — with a hard groundedness gate that blocks export unless every claim traces to a verified fact, plus ATS/recruiter simulation and dual-model QC. Built to make AI-written application material trustworthy, not hallucinated.

  • Groundedness verifier — a blocking gate: no export until every claim maps to a knowledge-base fact
  • ATS + recruiter simulation and knockout detection before drafting
  • Independent dual-model QC (Claude + Grok) with a consolidator
  • pgvector retrieval over a personal knowledge base; per-role multi-provider routing
Next.jsTypeScriptPostgres + pgvectorAnthropic + Google + xAIVercel AI SDK

Marketing Ops Console

Self-directed · runs my Shadow Edge marketing
Multi-agent marketing & CRM automation
Outcome
Runs a daily marketing loop end-to-end

A multi-agent marketing-operations system that runs a daily marketing loop end-to-end: campaign generation, compliance checks, consent-based lead capture, email-nurture queueing, distribution, and analytics reporting — with dry-run safety gates and human approval required before anything goes live. The CRM/marketing counterpart to RFP Factory: agents do the work, a human owns the go/no-go.

  • Orchestrates campaign generation, compliance review, and distribution as agent stages
  • Consent-based lead capture and email-nurture queueing with audit trails
  • Analytics import and daily operations reporting
  • Dry-run-by-default safety gates; nothing publishes without human approval
PythonMulti-agent orchestrationMarketing automationCRM / lead captureCompliance gating

Shadow Edge Tools

In production · shadowedgetools.com
NinjaTrader 8 add-on product line
Outcome
Designed, built, and sold end-to-end

A commercial product line of NinjaTrader 8 trading add-ons (Bracket Boss, Drawdown Guardian, with more in the pipeline) — owned end-to-end: the add-ons themselves, the storefront, checkout, and support. Full solo ownership from code to go-to-market, shipping real software to paying customers.

  • Built and shipped multiple NinjaTrader 8 trading add-ons as commercial products
  • Designed and launched the storefront, product catalog, and checkout flow
  • Owns the full go-to-market: pricing, licensing, fulfillment, and customer support
NinjaTrader 8 (NinjaScript / C#)Next.js 16React 19TypeScriptTailwind 4Lemon SqueezyVercel
Competency profile

Core competencies — with the receipts.

The radar maps core competencies across AI delivery and federal IT. Click any card for the receipts — actual outcomes from federal and commercial work.

Third-party signal

Awards, retention, references — the parts I can't self-assert.

Recognitions issued by institutions. Outcomes measured by the teams I led. References available upon request.

2008–2025

Public Trust — Background Investigation

U.S. Federal Government · Tier 4 (High Risk)

Held throughout 17-year SSA tenure. Federal background investigation; cleared to handle PII at the Social Security Administration. Eligible to reinstate.

2021

SSA Commissioner Award

Kilolo Kijakazi · Acting Commissioner, SSA

Awarded by SSA Acting Commissioner Kilolo Kijakazi for spearheading the emergency COVID-19 medical document upload system on MySSA — conceived, built, deployed, and adopted in weeks rather than the typical 12+ month enterprise cycle. Kept nationwide disability-claim processing alive during widespread office closures.

2020

FAC-P/PM-IT Certification (Lapsed)

Federal Acquisition Institute

Federal Acquisition Certification for Program & Project Managers — Information Technology (FAC-P/PM-IT): the federal government's standard credential for IT program leadership. Earned 2020; now lapsed.

2026

AWS Certified AI Practitioner

Amazon Web Services

AWS foundational certification covering AI/ML and generative-AI concepts, services, and responsible-AI practices on AWS.

Team outcome
100%
Direct-report retention

Twelve direct reports across three years as Branch Chief — zero attrition. Federal IT promotion is slot-based, not performance-based, so retention is the real signal: the team I built chose to stay through a period of high federal turnover.

Team outcome
4 of 4
Mentees promoted into PM leadership

Every project manager I mentored as a senior PM at SSA went on to run their own programs successfully — the career-development outcome the org structure made possible.

References
References available upon request
Battlefield to Boardroom

The leadership the rest of the field can't replicate.

A combat tour in Iraq with the 101st Airborne translated directly into how I run incident command, lead teams through high-stakes delivery, and show up for customers under pressure.

Military
Resilience under fire

Combat tour in Iraq with the 101st Airborne.

Delivery Leadership
Composure in high-stakes escalations

Owns the room when production is down and millions of users are affected.

Military
Scaling teams in chaos

Led infantry units through deployment cycles where adaptability decided outcomes.

Delivery Leadership
Federal IT leadership across 170 offices

12 direct reports (incl. 2 team leads), with IT governance and standards authority over a 340-technician field operation nationwide, serving 7,000+ Hearings Office employees.

Military
Mission-critical decisions

Trained to make irreversible calls with incomplete information, under time pressure.

Delivery Leadership
99.9% uptime SLA governance

Same decisiveness applied to incident command for regulated, 24/7 platforms.

Military
Doctrine + improvisation

Followed playbooks where they worked; rewrote them when reality demanded.

Delivery Leadership
ITIL discipline + pragmatic delivery

Operationalized ITIL where it created value, never as bureaucratic theater.

Military
Service to mission

Volunteered, served, sacrificed — and shows up the same way every shift since.

Delivery Leadership
Mission-driven service ownership

Treats every SLA, every customer outcome, every audit as a personal commitment.

1
Combat tour · Iraq
101st
Airborne Division
Insights

Perspectives shaping the next era of IT services.

Short, opinionated takes on leading AI delivery — how modern teams ship agentic systems that are grounded, gated, and trusted. Start with the featured long-form pieces below — or click into any short take.

Featured · Playbook

The First 90 Days

How I'd Run My First 90 Days Leading AI Delivery

The real risk in a new senior role isn't moving too slow — it's shipping fast on the wrong things. The first 90 days are about earning the standing to make the decisions that matter, then making them. Here's how I'd run them leading an AI delivery organization.

Read the full playbook
Featured · Essay

The model was never the hard part.

The rarest skill in AI product work is reading the workflow — fast.

The industry is obsessed with the model. Which one benchmarks highest this month, which framework to standardize on, how to squeeze a few more points out of a prompt. It's a comfortable obsession, because it's measurable. It's also mostly beside the point.

Read the essay

Every executive can write a strategy memo. Every engineer can write a function. The compounding leaders right now are the ones who can do both — hold the end-state in their head, write the agent graph that runs the business logic, sit in the customer call that exposes the gap, and then go fix it themselves before lunch. That hybrid wasn't possible a decade ago at any kind of scale; the tooling forced specialization. With multi-agent systems, modern dev environments, and AI-augmented engineering, the visionary builder is finally tractable. Hire for it.

Let's talk

Hiring for AI delivery leadership? Three ways to evaluate fit.

Paste your JD into the chatbot for a fit map · read the 90-day Playbook to see how Jordan would run the role · book a 15-minute intro below. Calendar is live, resumes are pre-tailored, chatbot is grounded in everything on this site.

Calendly
15-minute executive intro
Open in tab
Currently exploring

After 17 years at SSA, I moved into the private sector to build — Quadratic Digital is where I design and ship multi-agent AI systems hands-on. My focus is leading AI delivery as a hands-on player-coach: setting the architecture and quality bar and leading the team that ships agentic systems, where nearly a decade of federal delivery discipline and hands-on building compound. Platform fit over title.

Want a sense of how I'd run the role? Read the 90-day playbook
Where I want to play
AI delivery leadership (player-coach)

AI Delivery Lead, Senior AI Engineering Manager, or Director of AI Delivery — a hands-on player-coach who sets the architecture and quality bar and leads the delivery of multi-agent AI systems, reading the code and directing the work, not just approving it. Federal or commercial.

Federal / GovCon AI delivery leadership

Director or Practice Lead of Federal AI Services at primes and emerging federal-AI firms — where winning federal work AND leading multi-agent AI delivery is the job, and Public Trust eligibility plus a decade of federal delivery discipline are the moat.

Less of a fit

Pure individual-contributor roles with no team to lead · Roles where AI is a marketing veneer rather than real operational delivery · Pre-revenue seed-stage where federal credibility and delivery scale don't carry weight

Work arrangement
Based in York, PA · Open to remote · Open to in-office · Has led teams in both. Optimizing for the right job fit.
Federal eligibility
Public Trust · Tier 4 (High Risk) · Held throughout 17-year SSA tenure (2008–2025) — a Tier 4 federal background investigation (a suitability determination, not a security clearance). Eligible to reinstate within the standard reciprocity window.
Direct contact
Resume

What position are you looking to fill?

Or grab a universal version:Same résumé · ATS-optimizedSend-ready résumé (2 pages)