Over fifteen years of walking into complex environments and delivering — from the Pentagon to Fortune 500 boardrooms, from air-gapped SCIFs to cloud-native platforms. My career has spanned physical security, emergency communications, computer vision, confidential computing, ML training data, and real-time streaming infrastructure. Every pivot demanded mastering new technology fast and winning against entrenched competitors. Every one produced results.
At Confluent, I was responsible for the Army, COCOMS, and Fourth Estate practice — architecting Apache Kafka and Apache Flink solutions for DISA, Army AGUN, U-SIEM, and IPPS-A and many others. I designed and delivered a 45-connector JADC2 Common Operating Picture demo that became the company's flagship defense demonstration. Sadly, my position was eliminated in IBM's 800-employee RIF.
Before that: 170% of quota at Labelbox, multiple Project MAVEN awards at Clarifai for threat detection AI, FBI and NNSA wins at Fortanix for confidential computing, and $0 to $4M revenue building Honeywell's integrated security practice from scratch.
The numbers tell part of the story. The other part is this: customers trust me. Not because I tell them what they want to hear — because I don't. I tell them the truth about what our platform can and can't do, what the real architecture should look like, and where the landmines are, and work with them to architect for the best possible outcome for mission success today and in the future. That honesty turns customer relationships into lasting partnerships. Engineers call me when they have a problem because they know I'll give them a straight answer. Program managers pull me into meetings because I make their leadership confident. I don't bullshit — and in federal, that's rare enough to be a competitive advantage.
What I bring isn't just technical depth in Kafka, Flink, AI/ML, and defense architectures — it's the proven ability to learn any technology, engage any audience, and close business in the most demanding environments on earth. Eight years USAF, Secret clearance active.
I don't demo PowerPoint slides. I build fully operational systems — real streaming data, real dashboards, real architecture — and I put them in front of customers who can tell the difference. I'll spend nights and weekends building the demo that makes a general officer lean forward in their chair, because that's what wins programs. These are some of the systems I've built from scratch:
A fully operational Edge-to-Core JADC2 COP running on Confluent Platform with 45 Kafka connectors, 60+ topics, 7 simulated military data systems (DIA, MDA, CENTCOM, NGA, Coalition C2, DLA, FAA), ksqlDB real-time enrichment, Apache Flink stream processing, and MongoDB Atlas vector search — all Dockerized and deployable from a single command. EW/SIGINT intercept fusion, multi-domain aircraft and naval tracking, cyber threat correlation with MITRE ATT&CK, spectrum monitoring, and RAG-powered natural language query. The interactive dashboard below is a live representation — scroll down to explore it.
Confluent Platform, Apache Kafka, Flink, ksqlDB. Built 45-connector JADC2 demo, DISA captures, Army pipeline expansion. Sub-second latency at scale.
Computer vision, NLP, LLM pipelines. Project MAVEN threat detection at Clarifai ($6M+). Labelbox ML training data platform — 170% quota, $2.5M ARR.
Intel SGX-based data protection at Fortanix. FBI, NNSA, DOE wins. Air-gapped POC delivery. STIG compliance, Zero Trust architectures.
JADC2 Common Operating Pictures, EW/SIGINT fusion, multi-domain intelligence dashboards. Army, DISA, JAIC/CDAO, USSOCOM, DIA, NGA.
Honeywell integrated security ($0→$4M). Siemens 3D video analytics (2→16 countries). AtHoc emergency comms for Capitol Police & White House.
Disconnected, degraded, intermittent, limited — where most architectures fail. Kafka Cluster Linking, edge AI with ONNX, air-gapped deployments.
AWS GovCloud, Azure Government, C2S. Docker/Kubernetes orchestration. Enterprise systems architecture across DOJ and 48 sub-organizations.
15+ years closing strategic accounts. MEDDIC-disciplined. Equally effective in a flag officer's conference room and a developer's terminal. Every role produced revenue.
| Category | Primary | Secondary | Context |
|---|---|---|---|
| Streaming | Confluent Platform | Apache Kafka, ksqlDB | Multi-connector architectures |
Enterprise-grade Apache Kafka with built-in schema registry, RBAC, audit logging, and 200+ pre-built connectors. ksqlDB enables real-time stream processing with SQL syntax — turning raw sensor data into actionable intelligence without writing Java. Cluster Linking is the critical enabler for DDIL: it mirrors Kafka topics between edge, theater, and enterprise clusters so data survives even when satellite links drop. This is how sensor-to-decision pipelines stay alive across classification boundaries. | |||
| Compute | Apache Flink | ksqlDB streams | Real-time event processing |
Distributed stream processing handling complex event processing, windowed aggregations, and pattern detection across millions of events/sec. In military C2, Flink powers real-time threat correlation — joining SIGINT intercepts with aircraft tracks and EW emissions to produce fused intelligence. Exactly-once processing guarantees ensure no intelligence data is lost or duplicated, even during node failures or network partitions at the tactical edge. | |||
| Persistence | MongoDB Atlas | Vector search, triggers | AI-ready data store |
Document database purpose-built for operational intelligence. Atlas provides the persistence layer where processed streaming data lands for query, analysis, and AI/ML consumption. Native vector search enables semantic retrieval over historical signal data — an analyst can search "emissions similar to SA-21 acquisition radar" and get ranked results from millions of SIGINT records. Atlas triggers fire downstream workflows automatically when new threat data arrives. | |||
| Languages | Python | Java, SQL | Pipeline & integration |
Python drives data pipeline development, simulation systems, and AI/ML model integration — from Kafka producers generating synthetic military data to ONNX Runtime inference at the edge. Java powers production Kafka Streams applications and custom connectors. SQL through ksqlDB and Flink SQL enables rapid prototyping of stream processing logic that non-developers can understand and validate. | |||
| Infrastructure | Docker | Docker Compose | Multi-service orchestration |
Docker Compose orchestrates complete C2 demonstration environments with 30+ interconnected services — Kafka brokers, ksqlDB servers, Flink job managers, MongoDB instances, data generators, dashboards — all from a single command. This containerized approach mirrors real edge deployments: portable, reproducible, and deployable on disconnected hardware at FOBs, TOCs, or aboard ships. Every demo is a working proof of concept. | |||
| IoT | MQTT | ONNX Runtime, sensors | Edge sensor ingestion |
MQTT handles lightweight sensor telemetry from IoT devices — environmental monitors, perimeter sensors, HVAC systems, autonomous platforms — bridged into Kafka via source connectors for unified streaming. ONNX Runtime enables AI model inference directly at the edge without cloud connectivity: signal classification, anomaly detection, and threat identification running on tactical hardware. This is how AI works in DDIL — the model deploys forward with the data. | |||
| AI/ML | Sentence-Transformers | ONNX, embeddings | Semantic search & RAG |
Sentence-transformer models convert unstructured intelligence reports, SIGINT transcripts, and threat assessments into dense vector embeddings stored in MongoDB Atlas. This enables Retrieval-Augmented Generation (RAG) — analysts ask natural language questions and get answers grounded in actual collected intelligence, not hallucinations. Combined with Flink-powered enrichment, new intelligence is embedded and indexed within seconds of collection. | |||
| Domains | JADC2 / Military C2 | EW, SIGINT, Spectrum | Defense & Intelligence |
Joint All-Domain Command and Control — connecting every sensor and shooter across air, land, sea, space, and cyber. This stack implements that vision: aircraft tracks, naval positions, EW emitter detections, SIGINT intercepts, cyber threat indicators, and spectrum monitoring — all flowing through Kafka, processed by Flink, persisted in Atlas, and displayed on Common Operating Picture dashboards. Decisions at the speed of relevance, not hours-old PowerPoint slides. | |||
I'm actively exploring my next role — Senior Solutions Engineer, Pre-Sales Architect, or technical leadership in defense, public sector, cybersecurity, or data infrastructure. If you're building something that matters and need someone who delivers from day one, let's talk.
linkedin.com/in/ronmcneely →Authentication required to access resume data.