Mark H.

Nora Foundry is an independent architecture practice focused on AI workflows, data pipelines, and distributed systems. I work with teams moving from prototype to production - helping stabilize systems, reduce complexity, and make them reliable to operate at scale.

• Led architecture reviews and stabilization efforts for AI workflows and data pipeline systems moving from prototype to production
• Identified root causes of system instability across orchestration, state management, and data flow
• Designed event-driven and streaming architectures (Kafka, async systems) for production-scale workloads
• Improved reliability through better failure handling, observability, and workflow design
• Delivered clear architectural direction and stabilization roadmaps for teams under real production pressure

ShiftUp is an AI-driven intelligence platform focused on building data pipelines and workflow systems that connect external signals, structured data, and customer systems to drive decision-making.

• Built the stakeholder intelligence pipeline from scratch - ingestion, workflow orchestration, external data (SERP/APIs), and LLM-based analysis
• Designed systems to ingest, process, and operationalize large-scale data for real-time sales intelligence
• Implemented multi-stage data workflows (search → extract → synthesize) producing structured, actionable outputs
• Defined system boundaries, data flow patterns, and architecture for scalability, reliability, and future extension
• Featured as Lead Architect in ShiftUp announcement, recognized for designing event-driven data pipelines powering AI systems at Salesforce (See announcement: https://finance.yahoo.com/news/shiftup-assembles-world-class-technical-110000845.html)

Architect of event-driven systems, data platforms, and production-scale distributed software. At Salesforce, I spent over a decade designing and building fault-tolerant data pipelines and event-driven infrastructure powering Einstein Bots and other production systems. My work focused on systems where reliability, scale, and long-term architectural decisions matter - including pipelines processing billions of events per month in live production environments.

• Architected and built a Kafka-based event pipeline processing 10+ Billion events/month, supporting real-time chatbot workflows
• Designed and implemented fault-tolerant consumers (circuit breakers, bulkheads, retry queues, exponential backoff) for production reliability
• Built real-time analytics pipelines using Kafka Streams for chatbot performance and behavioral insights
• Early member of the Einstein Chatbots team, helping take the platform from 0 → large-scale production adoption
• Designed and implemented backend services (Java, Spring Boot) and integrations across Salesforce APIs (REST, Bulk, Apex, Flow)
• Contributed to team growth through hiring, mentoring, and establishing engineering practices

Desk.com was a customer support platform serving fast-growing businesses, where I led analytics and data engineering bridging the core application and real-time analytics systems.

• Built a real-time customer service analytics pipeline using Kinesis and distributed consumers to process high-volume event data
• Designed and implemented streaming data workflows to transform and aggregate support interactions into actionable insights
• Developed backend systems supporting large-scale request handling (360k+ requests/min), improving system throughput and reliability
• Built integration layers between application services and Elasticsearch for efficient query and data access patterns

LiquidPlanner (now Tempo.io) was a project management platform where I joined as the 5th engineering hire, building core product features and backend systems from early-stage through growth.

• Early engineering team member (5th hire), owning full-stack feature development from design through production
• Built end-to-end product features across frontend, backend, and data layers in a growing SaaS platform
• Designed and implemented PostgreSQL full-text search across core data models, enabling fast, scalable querying across application entities
• Open-sourced an “InfiniScroll” jQuery plugin (50+ GitHub stars), used to support scalable front-end data loading patterns

Built internal security and identity systems for AT&T’s Chief Security Office, supporting large-scale enterprise operations.

• Led a team of 5 engineers (database and web developers) building internal security and identity systems
• Designed and delivered identity and access management applications used by hundreds of thousands of AT&T employees
• Built web applications supporting enterprise-scale security workflows across authentication, authorization, and access control
• Owned system design and delivery across backend, database, and application layers in a high-trust environment

Peak Systems (later rebranded as UpTop) was a web and mobile development firm building consumer and enterprise applications, recognized for high-quality product and UX work.

• Built full-stack, n-tier web applications across database, backend, and frontend layers using ColdFusion and relational databases
• Delivered solutions for enterprise telecom (Qwest/CenturyLink) and startup clients (Skillshow)
• Contributed to production web and internal systems in a client-services environment with multiple concurrent projects

Built internal systems supporting operational workflows for the 787 program, improving efficiency and reducing processing time at global scale.

• Transformed a 7-stage paper-based onboarding workflow into a software-driven system for the 787 program
• Reduced processing time for global onboarding applications from 30+ days to under 5 days
• Built internal web applications supporting distributed teams and operational workflows across multiple regions
• Developed full-stack solutions across backend, database, and user interface layers

Early-stage R&D role focused on near-to-eye display systems, building software and automation infrastructure for image quality analysis and internal research workflows.

• Sole software engineer working with a small R&D team (including PhDs) to build an automated image quality lab
• Designed end-to-end systems spanning hardware interaction, data collection, processing, and reporting
• Built web-based tools to distribute and visualize test results across internal teams
• Co-inventor on 2 patents related to image processing and display systems