Marcelino Jackson

Experience

• Multi-cloud releases were slowed by a fragmented container build process. Consolidated it into a single multi-stage pipeline producing 12 Linux images, cutting image size by 40% and build time by 24%.
• Roadmap execution was inconsistent across a distributed engineering team. Introduced architecture reviews and code quality standards to operationalize GitOps delivery, improving on-time milestone delivery from 64% to 91%.
• Business users were blocked on developers for ad hoc SQL requests. Launched a self-service Snowflake Intelligence layer (Cortex Agents, Analyst, Search) handling 100+ daily NLP + chat queries and removing 18 developer hours/week from request queues.
• AISQL workflows had governance gaps that increased exposure risk in a production Snowflake tenant. Implemented RBAC and dual-control access policies, raising governed coverage to 96% and reducing policy exceptions by 57%.
• Profitability planning lacked machine-level forecasting by ship and voyage. Led a Snowflake ML + Cortex AI POC that achieved 86% forecast accuracy and improved pilot-route margin by 8%.

Experience

• Deployments across Azure DevOps and Snowflake were taking 1.8 days end-to-end. Rolled out CI/CD 2.0 with Policy-as-Code guardrails, cutting median lead time to 4.5 hours (90% faster).
• CI/CD 2.0 adoption was inconsistent across teams. Ran lunch-and-learns plus implementation guides that onboarded 16 engineers across 3 squads and reduced time-to-first-production deployment by 38%.
• Critical dbt/SQL compliance issues were escaping into late review stages. Implemented CortexAI multi-LLM governance with automated PR feedback, blocking 77% of critical non-compliant patterns pre-merge during POC validation.
• Sentiment modeling quality was limited and labeling effort was too manual. Fine-tuned Mistral 7B on anonymized clinical feedback with human-in-the-loop sampling, improving accuracy by 9% and reducing labeling effort by 30%.
• Patient issue trends were difficult to detect quickly in existing reporting. Deployed real-time interactive sentiment dashboards that helped leadership prioritize fixes and resolve critical patient-facing issues 40% faster.
• Code reviews were slowed by noisy semantic checks on dbt changes. Added RAG-based validation in CortexAI against business documentation, reducing false-positive review alerts by 52%.
• Security controls were manually rolled out, creating deployment friction and governance lag. Automated Snowflake RLS/CLS via Terraform CI/CD, achieving 100% policy-as-code coverage and cutting control rollout from about 2 days to 35 minutes.
• Build queues and runner saturation were constraining delivery speed. Scaled Azure DevOps agent pools using OpenShift and AWS-hosted runners, improving build reliability to 89% and increasing concurrent pipeline throughput by 60%.
• Python dependency resolution was adding avoidable delay to pipeline runs. Migrated dependency management to UV, improving overall build efficiency by 62%.
• Failed builds required frequent manual re-runs and extended recovery cycles. Added AIOps-driven self-healing feedback loops, reducing manual re-runs by 34% and lowering pipeline incident MTTR by 27%.

Experience

• Pipeline failures were discovered too late to prevent delivery delays. Built Azure DevOps observability dashboards with Python + REST APIs, cutting build-failure MTTD by 25% and increasing pipeline visibility by 35%.
• Regulated releases needed stronger control gates without slowing engineering output. Hardened build/release pipelines with DOE-compliant environment controls, sustaining a 100% audit pass rate and zero unauthorized production releases.
• Compliance approvals were slowing security changes and creating release risk. Served as the technical liaison between engineering and DOE compliance teams, reducing security-control approval cycle time from 4 weeks to 10 days (64% faster) without delaying releases.
• Secret detection was inconsistent across repositories and audit exposure remained high. Operationalized PR-triggered GitGuardian onboarding, reducing secret exposure incidents by 45% and lowering audit findings by 29%.
• Hybrid cloud patterns varied by division, creating uneven onboarding. Defined Azure/AWS architecture standards adopted by 3 internal divisions, reducing platform onboarding time for new teams by 35%.
• Legacy scheduler replacement options were unclear and template creation was slow. Ran Airflow, n8n, and Apache NiFi POCs with local-LLM-generated Terraform/Ansible scaffolds, cutting template authoring time by 45% and producing disposition paths for 100% of targeted workloads.
• TB-scale historical logs were difficult to analyze for early incident patterns. Architected an on-prem LLM pipeline (Ollama + LangGraph + RAG) to automate anomaly detection and accelerate incident triage.
• Load balancer and API gateway changes were manual and drift-prone across a pilot scope. Built a GitOps-backed Terraform platform for Citrix ADC and evaluated Kong API Gateway for DOE-compliant API traffic management, with a Streamlit self-service intake for 3 teams' repositories as a small-subset POC, eliminating manual config drift and reducing infrastructure change lead time from 2 days to 25 minutes (99% faster).
• Troubleshooting required hopping across disconnected operational tools. Unified observability signals across WordPress logging, PagerDuty, and Slack dashboards, reducing troubleshooting time from 3 hours to 25 minutes (86% faster).

Experience

• Delivery tooling differed across teams, increasing operational friction. Built and drove a DevSecOps modernization roadmap adopted by 6 major product teams, standardizing toolchains.
• Frequent rollback risk was limiting deployment confidence. Simplified branching and deployment workflows, increasing release frequency by 25% and reducing production rollback rate from 11% to 6% (45% reduction).
• Cross-division approvals slowed execution and destabilized sprint scope. Aligned agile delivery across business units, reducing leadership approval cycle time from 10 business days to 8 days (20% faster) and improving sprint scope stability by 30%.
• Cloud adoption stalled among legacy administrators and slowed IaC rollout. Led an AWS-first capability shift by mentoring 3 legacy system administrators, increasing IaC-driven change adoption from 28% to 76%.
• Data transformation deployments lacked a consistent release path. Engineered CI/CD-integrated ELT pipelines with AWS Fargate and Azure DevOps, standardizing dbt deployments.
• CDC orchestration across hybrid environments was brittle and hard to scale. Built resilient, parameter-driven Python orchestration patterns, sustaining 99.9% data synchronization accuracy across hybrid AWS and on-prem systems.
• Schema drift was leaking defects downstream into analytics consumers. Implemented automated data quality gates in Snowflake pipelines, detecting schema anomalies pre-production and reducing downstream data defects by 25%.
• Legacy CI tooling increased cost and slowed governance modernization. Led migration from Bamboo to GitHub Enterprise and HashiCorp Vault, completing transition 3 weeks ahead of schedule and reducing annual CI infrastructure costs by 12%.
• API traffic across transit systems relied on disparate, team-specific gateway configurations with no centralized governance. Leveraging previous experience deploying Kong at two startups, authored and presented a Kong adoption proposal and proof-of-concept for centralized API management, defining gateway-as-code practices, plugin standards, and a phased rollout plan for adoption by downstream product teams.

Experience

• PowerBI release processes relied on manual steps and expensive licensing paths. Built a custom PowerBI CI/CD framework via CLI and Microsoft Graph, avoiding $180K/year in premium licensing costs while cutting report onboarding time by 35%.
• Identity and integration friction was slowing analytics access. Standardized OKTA SSO and Snowflake-Salesforce integrations through Kong Gateway's OAuth2/OIDC plugins and governed dbt models, centralizing auth enforcement and reducing login issues by 28% and data sync latency by 30%.
• Snowflake environment setup was slow and drift-prone. Automated RBAC, warehouses, and schemas through Terraform-driven GitOps, reducing environment provisioning time by 45% and cutting configuration-drift incidents by 90%.
• Hybrid cloud data throughput and availability were below business targets. Co-authored the AWS/Azure hybrid data strategy across Snowflake, dbt Core, and Fivetran, increasing pipeline throughput by 26% and data availability by 15%.
• Vendor data ingestion was creating delays for downstream analytics. Streamlined ELT pipelines with Fivetran and Python-wrapped dbt workflows, accelerating external vendor-feed ingestion by 40%.
• CDC deployments required repeat custom work across environments. Designed parameterized orchestration templates, sustaining 99.9% execution reliability and reducing deployment rework by 32%.
• Critical data anomalies were reaching production dashboards before detection. Built SQL-based lineage and reconciliation validation suites, detecting 94% of critical data anomalies before they reached production dashboards.
• Freshness and schema-change impacts were hard to detect quickly. Implemented data observability across dbt models and Snowflake schemas, improving freshness transparency and reducing schema-change time-to-fix by 50%.
• Platform modernization funding required executive alignment and a clear roadmap. Authored and presented a 3-year DevSecOps strategy to the CIO and Board, securing modernization funding approval and pulling phase-one delivery forward by one quarter.
• CRM and data platform releases were colliding in delivery windows. Established the Salesforce DevOps Center operating cadence, aligning CRM releases with data platform transformations and reducing deployment conflicts by 20%.
• API traffic between cloud services and the hybrid data layer had no unified routing or throttling strategy. Architected Kong API Gateway as the central routing and governance tier between application services (AWS/Azure) and the Snowflake data platform, managing traffic policies with decK declarative configurations and enabling canary-style API versioning during data pipeline migrations.
• External vendor and partner APIs lacked centralized governance, creating inconsistent authentication and unpredictable latency. Deployed Kong API Gateway as the ingress layer for all vendor-facing and Salesforce integrations, implementing OAuth2 and rate-limiting plugins, request/response logging, and API lifecycle versioning — reducing integration failures by 32% and providing full API observability for the first time.

Experience

• Microservice delivery lacked a unified multi-cloud release model. Engineered Azure/AWS DevSecOps pipelines with Kong Gateway declarative config (decK) integrated into CI/CD, standardizing API deployment patterns, increasing release frequency by 20% and reducing change failure rates.
• Security checks were inconsistent and often late in the release cycle. Defined and enforced a Security-First deployment stance with automated SAST/DAST gates, reducing production security findings by 23% and accelerating approval cycles by 18%.
• Toolchain gaps and manual CI/CD steps were slowing teams. Developed custom Golang CLI utilities and mentored engineers on GitOps workflows, reducing manual intervention by 30% and shortening developer onboarding time by 25%.
• GitOps adoption was uneven across the organization. Established a Community of Practice and mentored 10+ developers to raise overall cloud maturity.
• Provisioning Kubernetes environments across clouds was slow and inconsistent. Architected a hybrid AKS/EKS/on-prem platform with Ansible and Terraform, reducing environment build times by 40% and enforcing configuration consistency.
• Resilience and data-operations practices were fragmented in critical stores. Institutionalized Chaos Engineering and DataOps for Snowflake and MongoDB, reducing defect escape rates and improving system recovery times by 21%.
• Microservices lacked a unified API gateway, exposing inconsistent routing, authentication, and rate-limiting. Selected and deployed Kong API Gateway (Enterprise) as the centralized ingress layer for the AKS/EKS platform, implementing rate limiting, JWT authentication, request transformation, and custom plugin development — reducing integration onboarding time for new services by 35%.
• Service-to-service communication across the hybrid Kubernetes footprint was ungoverned and difficult to trace. Implemented Kuma Service Mesh alongside Kong Gateway to enforce mTLS, enable distributed tracing via Kong Connect, and provide east-west traffic observability — reducing untracked internal API calls to near-zero and improving incident root-cause identification time by 28%.
• API lifecycle management was ad hoc, with no consistent deprecation, versioning, or consumer-notification process. Established an API lifecycle governance framework using Kong's developer portal, automated API documentation via Swagger/OpenAPI specs, and defined deprecation workflows — improving consumer adoption tracking and reducing breaking-change incidents by 40%.