Carlos Mariscal
Staff Software Engineer

Engineering leader. System architect. Team builder.

20 years turning complex problems into production-grade platforms. I design scalable, resilient backend systems in cloud environments, lead complex migrations, and mentor engineering teams to deliver measurable business impact.

Carlos Mariscal — Staff Software Engineer
220+ Support channels served by an AI assistant I designed and deployed
130M+ Customer records processed in sanctions screening architecture I led
26,000+ Structured review artifacts generated in an AI pipeline I built

About

Backend and platform engineer with deep experience building cloud-native systems, modernizing legacy services, and applying AI to solve real operational problems.

I've led and delivered systems across support automation, compliance platforms, subscription services, AI tooling, and defense applications. My work spans distributed systems, event-driven architectures, microservices, and engineering leadership. I'm especially drawn to roles where strong technical execution, architecture ownership, and pragmatic mentorship all matter.

With 20+ years spanning enterprise (Intuit, Zoox), defense (DoD satellite systems, naval battlefield simulators), and SaaS (GIS platforms), I bring a rare combination of deep hands-on implementation skill and the ability to lead architecture decisions across large organizations.

Java Spring Boot Python Go AWS Distributed Systems AI / LLMs Kafka Kubernetes Event-Driven Architecture Microservices Team Leadership

AWS Certifications

Validated cloud architecture and development expertise through AWS certification.

AWS Certified Solutions Architect Associate Badge

AWS Certified Solutions Architect

Designing distributed systems, selecting appropriate AWS services, and architecting resilient, cost-optimized cloud solutions.

Associate · Certified 2025
AWS Certified Developer Associate Badge

AWS Certified Developer

Developing, deploying, and debugging cloud-based applications using AWS services, SDKs, and CI/CD best practices.

Associate · Certified 2026

What Colleagues Say

Recommendations from engineers, managers, and leaders I've worked with.

"He created a tool that automated support across more than 220 channels — a game-changer, handling 85% of support message traffic and reducing support volume by 27%. Carlos is a natural leader who consistently elevates the quality of work around him."
Gurveer Khakh Software Engineer, Intuit · 3+ years together
"Carlos played a key role in refactoring legacy systems into scalable, modern architectures. He introduced performance optimizations across the oBill and Sanctions Screening platforms, many of which resulted in measurable revenue gains."
Raul I. Garcia Sr. Staff Software Engineer, eHealth · 7 years together at Intuit
"Carlos led the migration from on-premise to the cloud, transitioning legacy Mule services to modern Spring Boot microservices. He designed an event-driven architecture to tackle challenging idempotency issues — a critical advancement for the platform's reliability."
Tanvi Panchal Software Dev Manager, Ingram Micro · 6 years together at Intuit
"Carlos's leadership style stood out as exceptional — he encouraged me to explore new ideas, provided opportunities for growth, and helped me learn valuable lessons from mistakes without fear of judgment. His guidance had a lasting impact on my career."
Kittinan Ponkaew Software Engineer, Intuit · 4 years together
"He started making significant contributions within a week and within a month he was already leading a project with accolades from everyone he interacted with. He is the kind of person who makes the Manager's job easy because there is no need to manage him."
Rao Madhavrao Senior Director Bioinformatics · Former Manager
"As a Staff Engineer / Tech Lead, he has demonstrated technical expertise, accountability and leadership consistently. He played a crucial role in orchestrating the legacy services to modern microservices. He led a team which owned the Signup APIs with adept ownership and resolved many high priority & customer impacting issues. He has also mentored junior engineers and consistently nurtured and invested in their development."
Dheeraj Williams Murthaty Technical Program Manager, Ex-Deloitte · 2+ years together at Intuit

Selected Work

Systems I've architected and led that delivered meaningful scale, reliability, and business value. Click any card to view the architecture and details.

Intuit · 2023–2025

AI Support Assistant (AIDA)

Slack-based AI assistant using OpenAI, RAG, and internal knowledge sources that automated support across 220+ channels and reduced support volume by 27%.

+

Problem

Intuit's internal developer support operated across hundreds of Slack channels with high message volume. Support engineers were overwhelmed by repetitive questions, lacked visibility into support trends, and had no automated way to triage or respond to common inquiries.

Architecture

  • Slack App (Lambda + Go): Receives event messages, routes through the AIDA pipeline, updates threads with AI responses
  • AIDA Service (Java, Spring Boot, K8s): Core orchestration — manages conversation history, creates Jira tickets, triggers batch jobs, publishes metrics, integrates with OpenAI and Glean RAG
  • Support Eco Service (Go, Lambda): Generates StackOverflow-style Q&A from support interactions
  • Insights Pipeline (Python, Lambda, SQS, DynamoDB): Clusters support threads to surface doc gaps, enhancement opportunities, and training recommendations via specialized prompts
  • Metrics Dashboard (Java, Spring Boot, DynamoDB): Tracks escalation rates, bot accuracy, and support volumes
  • Retry Architecture: Dedicated retry service and events queue for transient failure handling

Outcomes

220+Support channels served
85%Message traffic handled automatically
27%Reduction in support volume
PythonJavaSpring Boot GoAWS LambdaDynamoDB SQSKubernetesOpenAI Glean RAGSlack APIJira API
Intuit · 2021–2023

Sanctions Screening Platform Migration

Led architecture and implementation for migration from Amber Road to LexisNexis — enabling compliant screening of 130M+ Credit Karma and 14M+ Mailchimp records.

+

Problem

Intuit needed to migrate its sanctions screening service after acquiring Credit Karma (130M+ users) and Mailchimp (14M+ users). The legacy Amber Road platform lacked reporting, batch screening, and audit capabilities needed for enterprise-scale regulatory compliance.

Architecture

  • Real-Time Screening API: Receives ScreenPartyRequest from product offerings, calls Bridger/LexisNexis, publishes results to Kafka ScreeningEventTopic
  • Batch Processing Pipeline: Multi-stage system — Batch Creation Processor, Persistent Batch Creation, SFTP Processor for S3 uploads, Batch Results Processor for completed screenings
  • Kafka Event Architecture: Multiple topics (SanctionScreeningTopic, SanctionScreeningEnvelopeTopic, ScreeningEventTopic) with dedicated consumers for Data Feeding, Event Processing, Alert Status, and Screening Status
  • Audit Data Warehouse: Stores Screening History, Contact History, Decision History, and Alert History for regulatory compliance
  • Data Migration: Built migration path from Amber Road database preserving historical data
  • Resilience: API retry for Bridger unavailability, Screening Event retry with exhaustion handling, error storage in Data Feeder Errors

Outcomes

130M+Credit Karma records screened
14M+Mailchimp records screened
FullReporting, batch, & audit capabilities delivered
JavaSpring BootKafka AWS EC2RDSS3 OpenSearchKubernetesArgoCD LexisNexis API
Zoox · 2025–Present

AI Performance Review Assistant

Designed an AI pipeline that generated 26,000+ structured review artifacts for 2,800+ employees, saving ~1,400 manager hours per cycle and reducing AI costs by 60–80%.

+

Problem

Zoox's biannual performance review process for 2,800+ employees required extensive manual effort from managers — collecting feedback, analyzing performance data, and writing structured reviews. The process consumed thousands of manager hours and lacked consistency.

Architecture

  • Automated Feedback Collection: System aggregates performance data, peer feedback, and self-assessments for each employee
  • AI-Powered Analysis: LLM-based processing generates structured review artifacts ensuring consistency and completeness
  • Parallelism & Batching: Cut end-to-end runtime substantially by introducing parallel processing and intelligent batching to eliminate bottlenecks
  • Intelligent Caching: Reduced AI API costs by 60–80% through caching strategies that avoid redundant LLM calls
  • Automated Access Control: Role-based access for generated artifacts, ensuring HR data sensitivity compliance

Outcomes

26,000+Structured review artifacts generated
2,800+Employees covered
~1,400 hrsManager hours saved per cycle
60–80%AI cost reduction via caching
PythonOpenAI API Prompt EngineeringCursor + Claude
Intuit · 2019–2021

Subscription Signup Idempotency Platform

Designed an event-driven idempotency solution with 7+ async consumers, tracking UI with replay capability, eliminating critical duplicate subscription bugs in QBO.

+

Problem

The Online Billing Platform suffered from critical production bugs: concurrent signup requests created multiple subscriptions, customer grants were inactivated on second-request failures, failure events published with incorrect correlation IDs, and poison messages went unprocessed. These directly impacted QuickBooks Online subscription success rates.

Architecture

  • Idempotency Check (DocumentDB): Signup Consumer upserts a SubscriptionSignupDocument keyed by webs_request_id. Sequence number determines new vs. retry vs. duplicate request
  • 7+ Async Consumers: Signup, Process Subscription, Subscription Caching, Event Publisher, Failure Event, Transaction History, and Signup Retry consumers — each publishing typed events to a SubscriptionSignupEvent queue
  • Event Sourcing: Every lifecycle event recorded in the document, creating a complete audit trail per signup request
  • Automatic Replay: When a request fails and a STANDBY duplicate exists, the system automatically replays the next queued request
  • Tracking UI & API: Front-end for PDs and L1/L2 Support to view signup events and replay failed transactions
  • Poison Message Handling: Retry consumer with configurable exhaustion — exhausted messages surfaced in tracking UI for manual intervention

Outcomes

EliminatedDuplicate subscription creation bugs
Real-timeSignup lifecycle visibility via tracking UI
AutomatedFailed transaction replay
5 engineersMentored in design patterns & distributed systems
JavaSpring BootDocumentDB ActiveMQRedisAurora DockerKubernetesArgoCD
Intuit · 2017–2019

Private Catalog Offers Refactoring & Migration

Refactored a 1,381-line monolithic service using SOLID principles, achieving 60% performance gains and 90% faster response times with Redis caching.

+

Problem

The Private Catalog Offers API was a monolithic 1,381-line service class with severe issues: database transitions called 1,236 times (53.9s for a single query), hundreds of redundant AddOn lookups, production worst-case response of 43,485ms, and 20-minute unit test runs. VisualVM profiling confirmed the bottlenecks.

Architecture

  • Strategy Pattern: Replaced monolithic conditionals with dedicated strategy classes — Addition, Migration, Change, and Switch query action strategies
  • Catalog Model Library: Extracted domain models into a reusable library for cross-service consistency
  • Redis/CouchBase Caching: AddOn criteria cached to eliminate redundant DB calls. Query criteria cached with scheduled refresh (2h/12h). Cache preloaded on startup.
  • Concurrent Processing: Parallelized getTransitions and getExtendedAttributes calls instead of sequential execution
  • Early Pagination: Moved pagination before processing (max 500) instead of after — drastically reduced data processed per request
  • Single Batch Query: Replaced 1,236 individual transition DB calls with a single batched query

Outcomes

60%Performance improvement via concurrency
90%Faster responses from Redis caching
1,236→1Transition DB calls reduced to single query
SOLIDStrategy pattern decomposition
JavaSpring BootRedis CouchBaseAWSAurora QuartzVisualVMJPA

Technical Skills

Core technologies and practices I work with regularly.

6 categories across languages, cloud, data, AI, architecture, and DevOps

Languages & Frameworks

JavaSpring Boot PythonFlask FastAPI Boto3Pandas Pydanticpytest Go TypeScriptSQL

Cloud & Infrastructure

AWSLambda S3DynamoDB AuroraEC2 SQS / SNSCloudFormation CDKKubernetes DockerArgoCD

Data & Messaging

KafkaActiveMQ RedisDocumentDB OpenSearchOracle SQL Server

AI & APIs

OpenAI APIClaude GeminiChatGPT CursorPrompt Engineering RAGAgentic AI GleanREST GraphQLSOAP

Architecture & Patterns

Distributed SystemsEvent-Driven MicroservicesDDD CQRSSAGA Pub-SubAPI Gateway Circuit BreakerThrottling IdempotencySOLID StrategyFacade CommandFactory

DevOps & Observability

CI/CDJenkinsBamboo ArgoCDGit Splunk Grafana VisualVMJUnit

Career Timeline

20+ years across enterprise, defense, and SaaS engineering.

2025 – Present
Senior Application Engineer
Zoox (Contract via Procom) · Foster City, CA
Led AI-powered performance review pipeline for 2,800+ employees. Partnered with HR and leadership on compliance. Reduced AI costs 60–80%.
2023 – 2025
Staff Software Engineer — Developer Success
Intuit · San Diego, CA
Designed AIDA AI support assistant across 220+ channels. Built metrics dashboard and insights pipeline.
2021 – 2023
Staff Software Engineer — Sanctions Screening Tech Lead
Intuit · San Diego, CA
Led Amber Road to LexisNexis migration. Architected event-driven compliance platform for 130M+ users.
2019 – 2021
Staff Software Engineer — Subscription Signup Tech Lead
Intuit · San Diego, CA
Designed idempotency solution with event tracking UI. Stabilized critical QBO signup APIs. Mentored team in design patterns.
2017 – 2019
Senior Software Engineer — Catalog Tech Lead
Intuit · San Diego, CA
Led Mule-to-Spring Boot migration. Achieved 60% performance gains and 90% faster cache responses via SOLID refactoring.
2010 – 2016
Senior Software Engineer / Tech Lead
Trey Software · San Diego, CA
Lead engineer for JIST — a DoD satellite resource management web application. Built rWaves cloud GIS SaaS platform.
2009 – 2010
Senior Application Developer
TechFlow · San Diego, CA
Designed business opportunity tracking system. Maintained GSA applications.
2005 – 2008
Senior Software Engineer / Systems Analyst
Predicate Logic · San Diego, CA
Built SOA integration for JIST-NET and IWPT. Oracle DBA and systems analysis for defense applications.
2000 – 2004
Application Developer
Sonalysts · San Diego, CA
Managed 7-person team building naval data repository for JSIMS battlefield simulator. Led 3-developer team on data warehouse.
1989 – 1999
Operations Specialist 1st Class (SW/AW)
United States Navy · San Diego, CA
USS Constellation Battle Group. JMCIS database, NTDS tactical systems, Link-11. Navy Achievement Medal.

Education & Military Service

Bachelor of Science

Chapman University · 1998

Computer Information Systems

United States Navy

Operations Specialist 1st Class (SW/AW) · 1989–1999

USS Constellation Battle Group. Managed JMCIS database and LAN connectivity. Supervised air and surface tactical warfare pictures using NTDS and Link-11. Navy Instructor at Fleet Combat Training Center Pacific.

Navy Achievement Medal Letter of Commendation

Reading List

Books that have shaped how I think about engineering, architecture, and leadership.

16 books on engineering, leadership, and innovation
After Steve: How Apple Became a Trillion-Dollar Company and Lost Its Soul
Tripp Mickle

Chronicles the diverging paths of Jony Ive and Tim Cook after Steve Jobs's death, revealing how Apple's shift from design-driven innovation to operational and financial optimization reshaped the company. A compelling study of how leadership transitions and cultural priorities determine whether a technology company innovates or optimizes.

Artificial Intelligence: A Guide for Thinking Humans
Melanie Mitchell

A clear-eyed exploration of what AI can and cannot actually do, cutting through both the hype and the fear. Mitchell examines deep learning, computer vision, and natural language processing with rigor, making it an essential read for engineers who want to separate real capability from marketing promises when building AI-powered systems.

Fundamentals of Software Architecture: An Engineering Approach
Mark Richards & Neal Ford

A modern, practical framework for thinking about architecture styles, component design, and the soft skills architects need to be effective. Covers trade-off analysis, architecture characteristics, and patterns like microservices and event-driven systems in a way that's immediately applicable to real-world decision-making.

Bring Your Whole Self to Work: How Vulnerability Unlocks Creativity, Connection, and Performance
Mike Robbins

Makes the case that authenticity and vulnerability are not weaknesses but leadership strengths that unlock trust, collaboration, and high performance. Directly relevant to building engineering cultures where psychological safety enables teams to take technical risks, raise concerns early, and mentor openly.

Where Good Ideas Come From: The Natural History of Innovation
Steven Johnson

Traces the recurring patterns behind breakthrough innovations—from adjacent possible and liquid networks to slow hunches and serendipity. Reframes innovation not as lone genius but as the product of connected environments, reinforcing why cross-functional collaboration and diverse technical exposure lead to better engineering solutions.

Drive: The Surprising Truth About What Motivates Us
Daniel H. Pink

Dismantles the carrot-and-stick model of motivation, showing that autonomy, mastery, and purpose are the real drivers of high performance. Essential reading for engineering leaders who want to understand why the best developers thrive on challenging problems, ownership of their work, and a clear sense of impact.

Steve Jobs
Walter Isaacson

The definitive biography of Apple's cofounder, built from over forty interviews with Jobs himself. An unflinching look at how relentless product vision, obsessive attention to detail, and willingness to challenge convention can reshape entire industries—while also exploring the human costs of that intensity.

The Dichotomy of Leadership: Balancing the Challenges of Extreme Ownership to Lead and Win
Jocko Willink & Leif Babin

Builds on the Extreme Ownership framework by exploring the tensions every leader must navigate—when to lead and when to follow, when to hold the line and when to be flexible. Draws from combat and business scenarios to show that effective leadership is never about absolutes but about finding the right balance under pressure.

Extreme Ownership: How U.S. Navy SEALs Lead and Win
Jocko Willink & Leif Babin

The foundational leadership book that argues there are no bad teams, only bad leaders. Distills combat-tested principles—own every outcome, keep plans simple, prioritize and execute—into a framework that translates directly to leading engineering teams through high-stakes deliverables and organizational complexity.

The Idea Factory: Bell Labs and the Great Age of American Innovation
Jon Gertner

Chronicles how Bell Labs became the most productive research institution in history by combining brilliant individuals with the right organizational structure, patient funding, and cross-disciplinary collaboration. A powerful case study in how engineering culture, physical proximity, and long-term thinking produce transformative breakthroughs.

Dealers of Lightning: Xerox PARC and the Dawn of the Computer Age
Michael A. Hiltzik

The inside story of Xerox PARC, where a small team of eccentric engineers invented the personal computer, laser printing, Ethernet, and the graphical user interface—only for Xerox to fumble the commercialization. A cautionary tale about what happens when visionary engineering is disconnected from organizational leadership and product strategy.

The Snowball: Warren Buffett and the Business of Life
Alice Schroeder

The most comprehensive biography of Warren Buffett, tracing how disciplined long-term thinking, compounding effort, and an obsession with understanding fundamentals built one of the greatest track records in history. A masterclass in patience, strategic focus, and the power of staying within your circle of competence.

Elon Musk: Tesla, SpaceX, and the Quest for a Fantastic Future
Ashlee Vance

An authorized look at how Musk simultaneously built Tesla and SpaceX by applying first-principles thinking, vertical integration, and an unrelenting pace of execution. Illustrates both the extraordinary outcomes and the brutal trade-offs of betting everything on engineering-driven disruption across multiple industries at once.

The Everything Store: Jeff Bezos and the Age of Amazon
Brad Stone

The definitive account of how Bezos built Amazon from an online bookstore into the infrastructure backbone of modern commerce. Reveals the relentless customer obsession, data-driven decision-making, and willingness to operate at a loss for years that made AWS, Prime, and the two-pizza team model possible.

iWoz: How I Invented the Personal Computer and Had Fun Along the Way
Steve Wozniak & Gina Smith

Wozniak's own account of designing the Apple I and Apple II from scratch, driven purely by the joy of engineering elegant solutions with minimal components. A reminder that the best technical work often comes from intrinsic curiosity, resourcefulness, and the drive to build something that simply didn't exist before.

The Innovators: How a Group of Hackers, Geniuses, and Geeks Created the Digital Revolution
Walter Isaacson

Traces the full arc of the digital revolution from Ada Lovelace through the internet, showing that every major breakthrough—the transistor, the microchip, the PC, the web—was the product of collaborative teams, not solo inventors. A compelling argument that sustained innovation requires both individual brilliance and the ability to work across disciplines.

Contact

The easiest way to reach me is by email or LinkedIn.

Hosting AWS · S3, CloudFront, Route 53, ACM
Ask me anything

Carlos's AI Assistant