Interview Prep · Private / Personal Use

Caylent — Senior Customer Solutions Architect

A working brief pulling the strongest evidence from my resume and LinkedIn history that maps directly onto the Caylent Sr. CSA job posting — role mission, requirement-by-requirement fit, the stories to lead with, honest gaps, and questions to ask them.
AWS Pre-Sales · 7+ Years $57M+ Pursuit TCV 14 Named Pursuits · 10 SOWs AWS DevOps Pro · SA Associate
The Mission
  • Partner with sales as a trusted AWS advisor to existing and prospective customers.
  • Work backwards from customer goals to determine and communicate solutions.
  • Collaborate with and enable AWS pursuit teams.
Your Assignment
  • Lead deep-dive architecture & design sessions; propose Well-Architected solutions.
  • Author proposals & SOWs capturing requirements and constraints.
  • Educate & evangelize AWS — blogs, white papers, webinars, presentations.
  • Win significantly complex pursuits; engage strategic stakeholders.
  • Mentor CSA peers; guide complex/strategic pursuits.
  • Advance team best practices and processes.
This job description is close to a paraphrase of what I've been doing inside Rackspace's Public Cloud Advisory practice since 2022: dozens of named AWS pursuits, $57M+ in combined opportunity value, many SOWs personally authored, deep-dive architecture sessions, Well-Architected reviews via MontyCloud, and peer coaching on WAFR/MAP/OLA delivery. The differences worth naming honestly: I've been doing this as an internal MSP/SI advisory consultant rather than inside an AWS ISV/consulting-partner org structured exactly like Caylent's, and my AWS certs are DevOps Professional + SA Associate rather than SA Professional. Everything else — the pre-sales motion, the SOW authorship, the mentoring, the evangelism (Substack, portfolio write-ups, live demos) — lines up directly.

If asked "why isn't more of this on AWS already?" — this is the honest answer, in order.

Why Multi-Cloud in the First Place

Building the same class of system — governance platforms, RAG pipelines, agentic app rationalization — on all three clouds was a deliberate test of what my certifications (AWS DevOps Professional, SA Associate, Cloud Practitioner, GCP ACE, Azure Fundamentals) actually translated into as working production judgment, not just exam knowledge.

Why So Much Landed on Azure

Rackspace's Public Cloud Advisory practice has been heavily weighted toward Azure engagements — that's client-driven deal flow, not a personal architecture preference. The EARE platform, the governance platform, Park Agent Chat, and the Content Pipeline all leaned Azure because that's where the client environments already were.

Where My Actual Preference Shows

3
AWS Certs
DevOps Pro · SA Associate · Cloud Practitioner
1
GCP Cert
Associate Cloud Engineer
1
Azure Cert
Fundamentals — entry level

AWS is the only cloud where I pushed to the Professional tier. And even on Azure-hosted systems, the model layer often still defaults to Bedrock — Park Agent Chat and the Content Pipeline both call Claude via Bedrock from Azure compute, and the from-scratch AWS builds (rarf-v4, EARE AWS, Advisory AI Governance) are Bedrock-native end to end. Not universal — Knowledge RAG and Digital Twin's Azure side use Azure OpenAI — but the pattern repeats enough to be a real signal.

What This Means for Caylent

I'm not walking in needing to learn Bedrock — I already have rarf-v4 (migration assessment accelerator), EARE AWS (regulated-industry app platform pattern), and the Advisory AI Governance Platform (compliance-as-code accelerator) as production-grade, Bedrock-native, reusable starting points. → Full AWS Solution Mapping

JD RequirementYour Evidence
7+ yrs architecting/building/operating on AWS Strong
AWS-centric advisory since 2022 at Onica/Rackspace (Bedrock, Lambda, Step Functions, CloudFormation, DynamoDB), plus AWS S3 migration work at SmarTek21 (2017+) and cloud/hybrid architecture roles going back to RSM (2015) and D+H (2013). 10+ years touching AWS and adjacent cloud infrastructure if you count the full arc.
10+ yrs infrastructure/app/db/network architecture Strong
30 years total, with architecture-level ownership at ERGOS, D+H, RSM, SmarTek21, and Rackspace — private cloud, hybrid infrastructure, Citrix platforms, RPA/DevOps infra, and full-stack AWS/Azure/GCP.
Assess & align business + technical strategy with customer needs Strong
Core of the Rackspace advisory role: discovery workshops, Capex vs. Opex TCO modeling, migration wave planning, and executive business cases across energy, healthcare, insurance, financial services.
Excellent oral/written communication; presentation & workshop leadership across SME → executive Strong
Author of 14 pursuit briefs and many SOWs; direct C-suite relationships at D+H and RSM; workshop facilitation across dozens of engagements; public writing on Substack and the portfolio site.
Expert understanding of cloud-native patterns & Well-Architected best practices Strong
Hands-on with the AWS Well-Architected Framework (6 pillars) and MontyCloud-driven WAFR engagements producing credit-maximizing remediation plans; serverless/IaC-first architecture on every recent build.
Ability to travel up to 25% Confirm
Not addressed in resume/LinkedIn — have a direct, comfortable answer ready (remote-based, travel-ready).
Preferred: HA / fault-tolerant architectures Strong
Multi-region/HA design experience embedded in Well-Architected reviews; digital-twin dependency-graph simulation work explicitly modeling failure/6R disposition scenarios.
Preferred: IT compliance frameworks (PCI, HIPAA, GDPR, security) Strong
HIPAA, SOX, FFIEC, OCC/FDIC exposure managing 20+ regulated bank/healthcare clients at D+H; built a governance platform scanning code against 10 frameworks including HIPAA, NIST AI RMF, ISO 42001, EU AI Act, SOC2, FedRAMP.
Preferred: pre-sales or internal strategic roadmap influence Strong
This is the central thread of the last 4 years — pre-sales solution architecture is the day job, not an adjacent skill.
Preferred: AWS SA Professional / DevOps Professional / Specialty certs Partial
Have: AWS DevOps Professional, AWS Solutions Architect Associate, AWS Cloud Practitioner. Gap: no SA Professional or Specialty cert yet — name this directly and pair it with the depth of hands-on WAFR/MAP/OLA delivery as the offsetting proof point.
Preferred: deep domain expertise in App Dev, Data, or DevOps Strong
App Dev: agentic AI systems shipped to production in days. Data: BigQuery/pgvector pipelines, forecasting (LightGBM, Prophet). DevOps: Scrum Master roles, CI/CD, IaC across three clouds.
Preferred: industry vertical expertise (FinTech, Healthcare, Oil & Gas) Strong
Financial services (D+H, RSM — OCC/FDIC/FFIEC regulated banks), Healthcare (clinical portfolio migration analysis, D+H healthcare clients), Energy (AWS strategic account engagement, datacenter-to-cloud TCO for energy sector customer).
Six ready-to-tell examples, picked for direct overlap with "deep dive architecture sessions," "win complex pursuits," "mentor peers," and "evangelize AWS."
$57M+ Pipeline, Dozens of Named Pursuits

Situation: Onica/Rackspace's Public Cloud Advisory practice needed a principal architect who could own the full pre-sales lifecycle across verticals.

Action: Personally authored many SOWs and led deep-dive architecture sessions across healthcare, financial services, telecom, retail, energy, and government.

Result: $57M+ combined opportunity value, 10 SOWs closed on 14 named pursuits.

→ 14 anonymized engagement briefs
WAFR, MAP, and OLA — Hands-On, Not Theoretical

Situation: Customers needed AWS partner-funded assessments to unlock credits and build migration business cases.

Action: Delivered WAFR via MontyCloud with credit-maximizing remediation plans, OLA via Evolve (Microsoft/SQL Server/Oracle licensing), MAP Assess via Flexera Cloudscape, and MAP Migrate/App Modernization via Tidal Cloud.

Result: Reusable solution patterns across 6+ verticals; direct peer coaching on program delivery.

→ Cloud advisory engagement detail
Built the Tool the Whole Team Uses

Situation: CSA peers were spending 4+ hours per account on manual pre-call research.

Action: Designed and shipped an agentic pre-sales intelligence tool producing 13-section company briefings in ~8 minutes from a single input.

Result: Deployed practice-wide, eliminating that research overhead for every CSA on the team — directly answers "raise everyone else's game," not just your own numbers.

→ Pre-Sales Intelligence Agent case study
36-App Clinical Portfolio Migration Analysis

Situation: A healthcare client needed a 408-VM, 36-application portfolio assessed for cloud migration disposition.

Action: Architected a multi-agent AWS Bedrock/Claude pipeline with 22 deterministic governance rules to eliminate hallucinated recommendations.

Result: Full assessment in under 5 hours for $24 in AI compute — a discovery timeline that used to take weeks, with provenance-tagged recommendations a customer's compliance team could trust.

→ Migration analysis platform architecture
20+ Regulated Clients, Zero Failed Audits

Situation: D+H's Citrix Private Cloud customers (banks + healthcare) faced multiple OCC/FDIC/FFIEC/HIPAA audit and exam events per year.

Action: Acted as virtual CIO / cloud services architect for the book of business, going on record for exam certification and remediation.

Result: $12M in renewal TCV over 18 months on a 20+ client, $4.5M book — directly relevant to Caylent's "compliance frameworks" preferred qualification.

Public Writing & Working Demos, Not Just Slides

Situation: JD asks for blogs, white papers, webinars, and direct evangelism of AWS.

Action: Independently run AI Pathfinder — architecture write-ups, live demos, and a Substack publication documenting real builds, costs, and trade-offs.

Result: 25+ published solution case studies with working code and architecture diagrams — proof this is a habit, not a one-off ask.

→ Full portfolio (25+ solutions)
Caylent is an AWS Premier Tier Partner (11x Partner of the Year, 850+ certs) organized around AWS Foundations & Migrations, Cloud-Native App Dev, Application Modernization, Generative AI, Data Modernization & Analytics, Infrastructure & DevOps Modernization, and their newest push — Caylent Accelerate™ for Agentic Cloud Operations (launched June 2026, built on Bedrock AgentCore, AWS MCP Server, and Amazon OpenSearch Serverless). Below, each area is paired with a specific system I've actually built — not a hypothetical — so I can go deep instead of describing patterns in the abstract.
1. Agentic AI on Bedrock — Guardrails, Provenance, Multi-Agent Orchestration
Why it matters: this is Caylent's newest flagship (Accelerate™) — guardrailed agents doing real work with human approval/audit, not just chatbots.
Enterprise COTS App Rationalization

Multi-agent pipeline on AWS Bedrock + Claude Sonnet analyzing a 36-app / 408-VM healthcare portfolio. 22 deterministic governance rules validate every agent output before it's trusted — the same "agents propose, rules/humans approve" pattern Accelerate™ uses for remediation. $24 total AI compute, 4.9-hour runtime, zero hallucinated numeric outputs.

→ Solution detail · Architecture
EARE Platform — Bedrock Multi-Agent Swarm + MCP Servers

5-agent Bedrock swarm plus 2 MCP servers running on ECS Fargate for a HIPAA-compliant healthcare platform generated entirely from spec. AWS MCP Server is literally one of the four pillars Caylent names for Accelerate™ — I've already shipped MCP servers in production, not just read the spec.

→ Solution detail · Architecture
Pre-Sales Intelligence Research Agent

Agentic tool-use system (Claude, live APIs) producing a 13-section company brief in under 10 minutes, replacing 4+ hours of manual research — deployed team-wide at Rackspace. Direct proof of "agents doing the highest-volume, lowest-differentiation work so humans focus elsewhere," Accelerate's core value prop.

→ Solution detail · Architecture
AI Governance Platform — SDLC Gate Enforcement

10-framework compliance scanner enforcing 5 human-approval gates before an AI system reaches production. This is my closest built analog to Accelerate's "guardrailed remediation" model — I already design the checkpoint/audit-trail layer that sits around autonomous agent actions.

→ Solution detail · Architecture
Talk track: "I haven't used Bedrock AgentCore or the AWS MCP Server specifically, but I've built the exact pattern it formalizes — guardrailed multi-agent pipelines with deterministic validation and human approval gates — on raw Bedrock and MCP servers on Fargate. I'd expect to be productive on AgentCore within days, not weeks."
2. Serverless-First Modernization — Lambda, Event-Driven, Zero-Fork Deployment
Why it matters: Caylent's Taco Bell case study leads with 90% infra cost reduction and 90% code complexity reduction moving to serverless.
Enterprise AI Governance Platform — AWS

Full Lambda + DynamoDB serverless deployment (5 CloudFormation stacks, 17 API routes) built as a cloud-adapter port of an existing Azure system — zero forked business logic. Direct evidence of designing serverless boundaries around existing application logic rather than a rewrite.

→ Solution detail · Architecture
Digital Twin App Dependency Simulator

AWS Lambda-based simulation engine running 5 AI simulations (wave rehearsal, blast radius, right-sizing, TCO) across 1,053 servers. Good example of choosing Lambda for bursty, event-driven analysis work rather than always-on compute.

→ Solution detail · Architecture
EARE Platform — Step Functions ETL with PHI Masking

Step Functions-orchestrated ETL pipeline with PHI masking as part of a HIPAA-compliant serverless architecture — shows judgment on where Fargate (long-running agent swarm) is the right call versus where Step Functions/Lambda (bounded, auditable steps) is correct.

→ Solution detail
Talk track: "My default bias is serverless for bursty or event-driven workloads — Lambda, Step Functions, DynamoDB — and I reserve ECS/Fargate for long-running agent swarms or anything with heavy VPC/networking requirements, which is exactly the boundary I drew in the EARE platform."
3. Data Modernization & Analytics — Warehouse Migration, Cost-Driven Architecture Decisions
Why it matters: Caylent's Criteria case study — legacy DW to a Redshift Serverless lakehouse in 4 months, 3x faster than prior efforts.
Database Architecture Decision: A Cost Inflection Story

Documented migration PostgreSQL → Cosmos DB → back to PostgreSQL after a hot-partition problem hit at 5.9M documents and the fix would've cost $394/month with no benefit. This is the single best story for a "walk me through a data platform migration decision" question — real numbers, real reversal, six documented lessons.

→ Solution detail · Architecture
ML-Based Weather Nowcasting System (BigQuery)

8 years of METAR + radiosonde data fused in BigQuery with a spatial boost via ST_DISTANCE; 6 scikit-learn models trained on 44K+ observations. Directly analogous to Redshift Serverless + Athena lakehouse patterns — translate BigQuery vocabulary to Redshift/Glue/Lake Formation on the fly.

→ Solution detail · Architecture
Operational Knowledge Retrieval System (pgvector)

ETL + SLM-routed query engine over 1.6M+ records with 3 query strategies (vector, numeric, hybrid) and deterministic fast paths that bypass AI inference entirely when not needed — relevant if the interview goes into vector search / RAG-in-a-lakehouse territory.

→ Solution detail · Architecture
Talk track: "I haven't personally migrated a warehouse onto Redshift Serverless, but I've made — and reversed — a real data-platform architecture decision under cost and performance pressure, with the numbers to back it up. The BigQuery pipelines are the same analytical pattern Redshift Serverless targets."
4. Migration Frameworks & Well-Architected — 6R, Dependency Mapping, TCO
Why it matters: this is the strongest overlap with your actual pre-sales background — WAFR, MAP, OLA — plus a working tool that operationalizes it.
Digital Twin App Dependency Simulator

Live 30-day telemetry collection across 1,053 servers / 4 datacenters, dependency graph with 174 edges, and an 8-rule 6R disposition engine producing wave-rehearsal, blast-radius, and TCO simulations. This is a working tool version of the exact WAFR/MAP discovery work described in the JD.

→ Solution detail · Architecture
Real AWS Program Delivery (resume, not demo)

WAFR via MontyCloud with credit-maximizing remediation plans; MAP Assess via Flexera Cloudscape; MAP Migrate & App Modernization via Tidal Cloud; OLA via Evolve for Microsoft/SQL Server/Oracle licensing. Dozens of pursuits, $57M+ combined TCV, many SOWs authored.

→ Advisory engagement detail · Pursuit library
Cross-Cloud Adapter Pattern (AWS ↔ Azure)

Same governance platform ported to AWS and Azure with zero forked business logic (Cosmos DB → DynamoDB, Key Vault → Secrets Manager, Azure AI → Bedrock, all invisible to handlers) — a concrete portability/reliability Well-Architected story with a diagram to point to.

→ Solution detail · Architecture
Talk track: "The Digital Twin is the tool version of the discovery work I've done manually on real MAP/WAFR engagements — dependency mapping, 6R disposition, blast-radius analysis, TCO — so I can talk about the framework and show a working model of it in the same conversation."
5. Cost & FinOps as the Universal Currency
Why it matters: every public Caylent case study leads with a number — 50% infra spend, 90% cost cut, 300% efficiency, 65 days → 3 days.
$24 / 4.9 hours — Clinical Portfolio Analysis

36 apps, 408 VMs, full migration disposition analysis for $24 in AI compute — a discovery timeline that used to take weeks of manual work.

→ Solution detail
$394/mo → $0 — Cosmos DB Cost Reversal

Identified a hot-partition problem whose "fix" cost $394/month with zero benefit, migrated back to PostgreSQL, and documented six lessons — a real cost-optimization reversal, not just a savings number.

→ Solution detail
<$100/mo — Entire Multi-Cloud Platform

25+ deployed solutions across AWS, Azure, and GCP running under $100/month combined — a defensible answer if asked to reason about running cost-efficient production systems at small scale before they justify Reserved/Savings Plans conversations.

→ Platform overview
4+ Hours → Under 10 Minutes — Time as Cost

Pre-sales intelligence agent eliminated 4+ hours of manual research per account, deployed team-wide — the same "engineer time reclaimed" framing Caylent uses for Accelerate's remediation-time savings.

→ Solution detail
Talk track: "I default to quantifying outcomes the same way Caylent's case studies do — a dollar figure or a time figure, not just 'it got faster.' Every project I'd bring up has a number attached because that's what makes a technical decision defensible to a customer's CFO, not just their engineering team."
If asked to design something live, default to narrating the closest system you've actually built, then generalize outward — it's more credible than a purely hypothetical design.
"Design a customer-facing GenAI agent for a regulated industry (healthcare or financial services)."
Lead with the EARE Platform: Bedrock multi-agent swarm + MCP servers on ECS Fargate, Step Functions ETL with PHI masking, HIPAA-compliant VNet/networking, zero hand-written IaC. Cover: RAG/knowledge grounding, PII/PHI masking before it reaches the model, guardrails, and human escalation paths. → eare-aws.html
"Decompose a monolith into an event-driven serverless architecture."
Lead with the Enterprise AI Governance Platform — AWS: Lambda + DynamoDB + CloudFormation, ported from an existing Azure system via a cloud-adapter layer with zero forked business logic. Cover: where Lambda ends and Fargate begins (long-running agent swarms), and the adapter-pattern approach to avoiding a rewrite. → advisory-aws.html
"Walk through a legacy data warehouse → lakehouse modernization."
Lead with the Database Architecture Decision cost-inflection story (Cosmos DB → PostgreSQL) for migration reasoning and risk, then bridge to the Weather ML Pipeline's BigQuery analytics fusion as the lakehouse-equivalent pattern (translate to Redshift Serverless + Glue + Lake Formation live). → data-pipeline-evolution.html · → weather-ml-pipeline.html
"Design an agentic cloud-ops system that detects, diagnoses, and remediates issues autonomously."
No exact 1:1 match — be honest about that — but bridge through the AI Governance Platform's 5-gate SDLC enforcement model (agents/automation propose, deterministic rules and humans approve at defined checkpoints) and the rarf-v4 22-rule provenance validation layer. Both are the "guardrails around autonomous action" pattern Accelerate™ formalizes. → governance-platform.html
"How would you structure a WAFR engagement for a customer with a distributed team, given the role's ~25% travel expectation?"
Tests delivery-model fluency, not architecture. Reference real WAFR delivery via MontyCloud (resume) plus remote-first workshop facilitation across dozens of Rackspace engagements — most discovery was already done remotely, with travel reserved for executive workshops or contract milestones.
Condensed from the full AWS Solution Mapping page — every Azure/GCP-built project translated to its AWS-native equivalent, plus the container/serverless and database lessons learned along the way.
AWS Solutions Architect Professional certification

Currently hold DevOps Professional + SA Associate, not SA Professional or a Specialty cert. If asked directly: acknowledge it, then pivot to the depth of hands-on delivery (WAFR, MAP, OLA, multi-region Well-Architected reviews) as evidence the practical bar is already met — and state intent to sit for SA Pro if it matters for the role.

Formal AWS consulting-partner org experience

Rackspace/Onica is an AWS partner, but the org structure differs from a pure-play AWS consultancy like Caylent. Frame it as: same motion (pursuit team collaboration, SOW authorship, AWS program delivery), different org wrapper — and that the muscle memory transfers directly.

Travel expectation (up to 25%)

Not addressed anywhere in resume/LinkedIn. Have a direct, confident answer ready before the call — don't let this surface as a surprise.