Fred W. Nix Jr. | Principal Solutions Engineer | Infrastructure and AI Platforms
Most recent role · April 2026 to October 2026
Senior Solutions Architect, BlueAlly
Pre-sales and post-sales solutions architect supporting six account executives across the Southeast: enterprise, healthcare, higher ed, state and local government, and nonprofits. I owned discovery, architecture, sizing, bills of materials, cost and return models, proposals, proof of concept design, demos, and custom code on 24 engagements in five months. Roughly $150K closed in the first 90 days, a seven figure qualified pipeline. Customers are anonymized; the numbers are real.
01 · Infrastructure & data center
The platform work
- VMware exit, three continents. Global corporate travel company with production DCs in Atlanta, Frankfurt, and Singapore. Discovery, hosted POC, licensing strategy, and a parallel three site migration plan against staggered Broadcom renewals. Stretched storage over NVMe/TCP on all flash arrays avoids ~$500K of Fibre Channel fabric.
- 190 property hotel chain. 480 core VMware entitlement review that closed as a signed $95K order in the first 90 days, legacy SAN retirement (34 TB), flash array capacity planning at 91% utilization, backup and DR discovery, five year TCO. Storage modernization follow on in flight at ~$210K gross profit.
- 18 MW immersion cooled colo. Multi vendor server, storage, and firewall BOM for a 250,000 sq ft, three hall operator with a GPU hosting relationship behind it.
- HCI to disaggregated. Five node hyperconverged cluster to two socket servers, an all flash NVMe array, and private cloud automation for an Atlanta media ministry, staying on VMware by design.
- 215 TB video archive. Storage discovery for a Florida state water district: 130 cameras, four sites, commercial VMS, tiering and protocol design.
- TCO and RFPs. 48 host virtualization break even model for a materials manufacturer, hypervisor migration evaluation for an aerospace and defense contractor, all flash storage RFP for a major international airport, SIP trunking and SBC eRFP for a state research university with BlueAlly as accountable prime.
02 · Security & identity
The trust work
- Cloud RADIUS and EAP-TLS at 100 sites. Multi state HIPAA healthcare provider, 600+ daily users. Retire on prem NPS and 100 branch VPN tunnels, move wireless auth to cloud RADIUS with certificate lifecycle delivered as a service on the cloud managed wireless estate they already own.
- Global IAM. Agriculture and food company, 150,000 employees, 77 countries, ~2,000 legacy apps. Identity as the door into Microsoft E5 posture, secure service edge, badge and logical access convergence, AI governance.
- Contract compliance save. Caught a managed services decommission that would have dropped monitored assets below contracted minimums at a higher ed account. Escalated, preserved revenue and standing.
- Zero trust design patterns across every engagement: air gapped AI, PCI scoped hospitality, HIPAA scoped healthcare, state entity procurement rules.
03 · Private AI & GenAI
The AI work
- On prem GenAI reference architecture. 2U four GPU node, 4x 96 GB GPUs (384 GB per node), private AI appliance stack, orchestration and governance layers. Sized to ~14M peak monthly tokens. Sold as a fixed price 60 to 90 day POC to decouple from CIO budget gating.
- Nine use case portfolio, $12.1M. AZA accredited research institution. 17,800 reclaimed hours (8.5 FTE). appliance sized from 109 TB to 343 TB of KV retention. Five year lease above $2M. Data sovereignty as the wedge against cloud only competitors.
- Academic medical center, $2.4M. Six prioritized use cases across clinical documentation, research ops, student success, and reporting. ~1,300 reclaimed hours.
- Contract intelligence agents. Healthcare payments company: platform vs custom build evaluation, pass through delivery model, $290K to $305K proposal.
- Workshops and assessments. Office REIT ($6.7M perceived value, Copilot Studio vs agent platform on Fabric), commercial HVAC platform (1,200 employees, 33 sites), county government, state financial regulator, retirement services firm, EPC engineering firm (signed AI readiness and Copilot Studio SOWs, ~$55K), youth services nonprofit.
- OEM partner funding. Own the SOW for a $62.5K OEM demand generation program: a customer discovery workshop plus AI certification training for the BlueAlly sales and SA team.
Reference architecture · click a layer
The private AI stack I design and use daily
The layered model behind the 2026 engagements. Sovereign by default: no cloud, no API keys, data never leaves the customer's network.
L6 Applications & agents Custom apps · Copilot Studio · contract intelligence · clinical documentation L5 Governance & security policy engine · RBAC · audit · PHI, PII, IACUC scoped data paths · air gap option L4 Orchestration RAG pipelines · MCP servers · OpenAI compatible endpoints L3 Models & serving vLLM · LLaMA, Mistral, fine tunes · AWQ/GPTQ · tensor parallel L2 Compute & GPU 2U four GPU node · 4x 96 GB GPUs · 384 GB GPU memory per node · blade chassis clusters · 48 GB GPUs L1 Storage & network All flash NVMe arrays · NVMe/TCP · 100GbE · 109 to 343 TB KV retention
Previous role · March 2025 to April 2026
AI Solutions Architect, Business Technology Architects
Storage and Compute Practice Manager. Built a production grade, air gapped AI inference platform for channel partners on modular blade compute, 48 GB GPU clusters, all flash storage, and 100GbE.
100%
deployment success rate. One command from bare metal to running inference across every partner rollout.
Multi model
LLaMA, Mistral, and custom fine tunes with hot swap switching, vLLM, AWQ and GPTQ quantization, tensor parallelism.
OpenAI API
Compatible endpoint plus a ChatGPT style UI with streaming and live GPU metrics (DCGM, Prometheus, Grafana) for LangChain and LlamaIndex RAG.
Enablement
Trained channel partners with hands on labs and documentation built for repeatable success.
Skills · infrastructure first
What I bring to the whiteboard
Depth by domain
Infrastructure 96 Storage 95 Virtualization 86 AI / GPU 88 Presales 97 Automation 80 Security 76
Infrastructure
Two socket and four GPU rack servers · modular blade compute · all flash NVMe arrays · scale out NAS · VMware vSphere & VCF · hyperconverged and disaggregated designs · private cloud automation · Proxmox · NVMe/TCP · 100GbE · SD-WAN · backup & DR · colocation & immersion cooling · VDI
AI & GPU
GPU sizing from 48 GB and 96 GB cards to 384 GB nodes and multi node clusters · private AI appliance stacks · vLLM · AWQ/GPTQ quantization · tensor parallelism · RAG architecture · OpenAI compatible APIs · LangChain & LlamaIndex · Copilot Studio · MCP servers · Ollama local inference
Presales & business
Discovery workshops · use case portfolios and value modeling · TCO, ROI, break even · BOM · RFP and eRFP responses · POC design · lease and pass through structures · OEM partner funding programs · executive briefings
Development & automation
TypeScript & Bun · Python · Claude Code and Personal AI Infrastructure · git based knowledge systems · Cloudflare Pages & Workers · Linux (Omarchy, Hyprland) plugin author · Prometheus & Grafana · GitOps
Security
Zero trust access · EAP-TLS & cloud RADIUS · air gapped deployment · HIPAA and PCI scoped design · identity & access management · AI governance
AI augmented practice · since January 2026
How I build: an agentic engineering platform
A four host Claude Code platform I built and run every day: version controlled memory, written operating rules the agents follow, local first model routing, and hooks that enforce code quality. It is how a customer conversation becomes a demo, a sizing, and a deck in about a day and a half.
Fleet · hover a node
gus 12 GB GPU · NPU vic 6 GB GPU fnix ARM · unified memory blumemory tunnel memory git + qdrant
Four hosts, one identity. Memory is git plus a vector index, so every architecture decision and rejected approach is searchable months later.
Model router · live simulation
Every prompt is classified and sent down a cost ladder: local GPU first, cloud only when it earns it. Type something and watch where it lands.
Local GPU Ollama · 12 GB GPU $0.00
Haikufast, cheap ~ $0.001
Sonnet balanced ~ $0.015
Opus architecture, code that ships ~ $0.075
Waiting for input.
4 hosts one identity, host local memory
Local first a 12 GB local GPU before any cloud call
Every session indexed into searchable memory
10+ sites and 30+ desktop plugins shipped
Selected AI engineering projects
Things I built and shipped
Five running systems. Full write ups, including the AI audit trail and the software factory, at resume.nixfred.com/projects.
01 · Agentic engineering platform
A coding setup that remembers
Four host Claude Code platform with MCP integrations and local first model routing (a 12 GB local GPU before any cloud call). Persistent agent memory in TypeScript and Bun with SQLite full text plus vector search, lifecycle hooks, a written constitution of operating rules, and scheduled recovery of missed session extraction.
02 · AI software factory
Spec in, live site out
A reusable AI assisted delivery workflow in TypeScript, Bun, Python, Cloudflare Workers and Pages, and GitOps. Ten plus production sites and 30 plus desktop plugins shipped; the same tooling turns a customer conversation into a demo, a sizing, and a deck in about a day and a half.
03 · AI Systems Workbench
Fourteen on device instruments
Browser based tools for designing and evaluating AI applications with no server behind them: prompt lab, token and cost planner, retrieval lab, evaluation workbench, agent designer, least privilege permission planner, latency budgeter, failure investigator, drift monitor, AI stack mapper.
04 · AI audit trail and governance
Agents need a system of record
Every agent session is committed to git with its transcript, decisions, and rejected approaches, then indexed for search, so any answer can be traced months later. Least privilege permissions and an evaluation workbench for critical failures, coverage gaps, and drift.
05 · Open source contributor
Linux desktop, custom development
30 plus published desktop plugins for Linux (Omarchy, Arch based) and 120 plus public repositories: hardware control, media, messaging bridges, and system monitoring.
Career · 2010 to 2026
Twenty years of enterprise infrastructure
2026
Senior Solutions Architect
BlueAlly Technology Solutions, Atlanta · Apr to Oct 2026
Infrastructure presales architect. 24 engagements: VMware exits, storage and data center refreshes, identity, plus the private AI reference architecture and $21M+ in modeled AI value.
2025
AI Solutions Architect, Storage & Compute Practice Manager
Business Technology Architects, Alpharetta
Air gapped AI inference platform for channel partners. One command deploy, 100% success rate.
2023
Partner, VP Sales
EMOTO Electric Motorbikes
Sales up 225% in year one. Scaled to regional market leader.
2021
Enterprise Sales Engineer
Pure Storage
Fortune 100 accounts, Southeast. 250% of quota in 2021, named Global Rookie of the Year. 100% uptime for a customer through a hurricane disaster recovery event.
2019
Solutions Architect, Converged Systems
Dell Technologies
Software defined storage and cloud presales. Major data center consolidations.
2015
Global Presales Director
Reduxio
First American employee, led US market entry. 150%+ of quota annually, 90%+ conversion on 50+ POCs.
2010
Global Presales Operations & Technical Enablement
EMC Corporation
Operations for 5,500 presales engineers worldwide. Trained 1,000+. Grew vSpecialists from 30 to 150+.
Earlier
CTO · SC · PS
Senior technical roles
Syscom, MTI, GT Software, Vitria, Windward, Kilpatrick Stockton, Medaphis
CTO, solutions consulting, presales, professional services, and IT administration.
Education & service
B.S. Business Administration, Information Systems
Auburn University · U.S. Army veteran
Track record
Annual quota achieved every year of my career
Six account executives supported in 2026. Pre-sales, post-sales, demos, custom code, and the business case.