Speakers

BIO: A Skyspark power user for over a decade, I started in the world of system integration becoming fluent in axon and general database management, then moved into lower level fantom, eventualy contributing to both the open source Project Haystack and Haxall projects. Now diving deep into the new era of xeto and looking forward to all the doors this new technology will unlock!

ABSTRACT: Learn how to leverage the widely used Swagger tool chain with SkySpark's new HTTP 5 APIs including auto-code generation from Xeto function specs. See how easy it is to create new Swagger endpoints with a simple Xeto function spec.

ABSTRACT: SkySpark ontologies usually get built from the BAS side: point a connector at the building and tag whatever the learn brings back. The design drawings, which say what the systems actually are, sit in a folder the whole time. We flipped that. We built a pipeline that extracts the ontology from the drawings themselves, so the model exists before the first byte of telemetry arrives. This talk is about how that pipeline is put together, stage by stage, and what each stage turned out to need.

Classification comes first: each page gets tagged by discipline and page type, then routed to a specialist extraction prompt, since a mechanical schedule and a panel schedule are different reading problems. The output is candidate equipment records, each field traced back to the page it came from. An agent then assembles those candidates into equip and system records against our Xeto specs, and everything it proposes gets reviewed and committed through the same path as a manual edit.

The back half of the session is what we learned running this on real projects: how we keep the extraction prompts under control, where the models go wrong, and whether the hours saved justify building any of this. We'll bring the numbers.

BIO: A key player in the founding of Project Haystack, Jason Briggs is a visionary entrepreneur and passionate innovator. At just 21 years old, Jason co-founded Controlworks, a systems integration company recognized for completing one of the largest building automation system (BAS) installations in Las Vegas at the time. He went on to found J2 Innovations in 2009. Under his leadership, the company developed the acclaimed product, Fin Stack, which has played a key role in the growth and success of the company.

Today, Jason continues his entrepreneurial journey as a proud founding member of Novo Studio, a venture studio that specializes in nurturing and supporting startups in the smart building space. He is excited to make a lasting impact in the industry and help other entrepreneurs achieve similar success.

ABSTRACT: Explore the new cloud repository for Xeto libs and documentation - the new centralized hub for the Haystack ecosystem. Xeto.dev will be used crowd source specs, equip templates, AI skills, Axon functions, and visualizations.

BIO: Joining LBNL in 2019, Armando is an Energy Technology Researcher working on several projects related to building energy efficiency, model-based design, demand flexibility, building commissioning and fault detection and diagnostics. Armando worked as an Energy Analyst at UC Merced Facilities Management and has also worked for the Energy Conservation Office at UC Davis performing continuous commissioning. His Masters work was conducted at the Western Cooling Efficiency Center, which involved wireless sensing for building leakage diagnostic applications. Armando holds a M.S in Energy Systems from UC Davis and a B.S in Mechanical Engineering from UC Merced.

ABSTRACT: Utilities across the United States are increasingly turning to demand side management to address rapidly growing network demand. Buildings represent a major flexibility resource, capable of supporting demand response and other grid services. However, retrofitting buildings with demand flexible control sequences remains difficult, as no widely adopted specification exists for demand flexibility comparable to ASHRAE Guideline 36 for efficiency. To address this gap, we are developing a set of demand flexible control strategies and implementing them using the ASHRAE S231 Control Description Language (ASHRAE S231). A key part of this effort is ensuring these sequences are portable across buildings, which we achieve by leveraging Xeto to capture relevant constraints. In this talk, we demonstrate how we develop and use these demand flexible strategies, how we implement them using ASHRAE S231 sequences, integrate Xeto, and use these Xeto specifications to validate building models and connect them to building systems.

ABSTRACT: Onboarding a new site into SkySpark is one of the most time-consuming and error-prone steps in any deployment—teams manually translate raw source documents into build files, then hope nothing was missed or mistyped along the way. This presentation walks through a working, AI-assisted pipeline that closes that gap end to end, turning raw documentation into a validated, build-ready site.

BIO: Brian Frank is the President and a Co-Founder of SkyFoundry, a software company specializing in the storage, analysis, and visualization of data from the Internet of Things. SkyFoundry’s software is used across the world in thousands of sites to analyze and optimize IoT systems including energy, HVAC, and lighting. Previously, Brian was a founder of Tridium and architect of Tridium’s software framework products. He designed and implemented the core subsystems for two generations of the Niagara Framework, widely acclaimed as a leading technology in the IoT space. During his tenure at Tridium, he also led the design of the open source Sedona Framework—a programmable control engine for smart sensors and embedded controllers. Brian currently serves as the technical lead for Project Haystack. He works with the Project Haystack community to curate domain models and technical specifications.

BIO: Matthew Giannini is a software engineer with close to 18 years of experience writing software in the BAS industry. His current role is software engineer at SkyFoundry, where he works on the SkySpark analytics platform. In that role, Matthew had the opportunity to become more involved in the open-source community, where he is also a contributor to the Fantom programming language and an active member of the Project-Haystack community.

ABSTRACT: Review how the 4.0 rule engine leverages Xeto specs and improves point binding. We will cover the details for the new rule silence feature. And discuss the new hx.comps function block library and upcoming block engine.

BIO: Dr. Jessica Granderson is a Senior Scientist and Director of the Building and Industrial Energy Systems Division at the Lawrence Berkeley National Laboratory. She is a member of the Whole Building Systems Department. Dr. Granderson holds a PhD in Mechanical Engineering from UC Berkeley, and an AB in Mechanical Engineering from Harvard University. Her research focuses on demand flexibility, energy management and information systems (EMIS), analytics and fault diagnostics, and advanced control. She is the recipient of the 2015 Clean Energy Education and Empowerment (C3E) Award for Leadership in Research, the 2020 Federal Laboratory Consortium Award for Excellence in Technology Transfer, and the 2021 Building Commissioning Association Benner Award (jointly).

ABSTRACT: The benefits of existing building commissioning are well known, and include enhanced energy efficiency and safety, lower operation and maintenance costs. Despite proven energy savings of 5-10%, and payback periods of one year, existing building commissioning remains underutilized. This talk will present how SkySpark and related technologies are being used to conduct automated commissioning (ACx) and functional performance testing. It will share recent findings from Lawrence Berkeley National Laboratory that quantify the value proposition for use of ACx, and insights from providers who use it. Further, Dr. Granderson will discuss the publication of open specifications to implement ACx in common building HVAC systems. The talk will conclude with an overview of ongoing work to document the use of ACx in the field, including a case study of how Otto is being deployed in a large university retro-commissioning project.

ABSTRACT: Air handling unit reset strategies are supposed to save energy, but in hospital environments they often backfire. Faulty terminal units send bad signals to the BAS, and reset logic responds by driving AHUs into inefficient, unstable operation. Most analytics platforms can flag the problem. Few can fix it.

ENFRA, in collaboration with Lawrence Berkeley National Laboratory, built a SkySpark-based analytics layer that identifies rogue zones in real time, isolates their influence on reset calculations, and writes corrected setpoints back to the BAS. This session walks through the Haystack tagging model, the rule logic used to detect and suppress rogue zone influence, and the write-back architecture that lets SkySpark move from passive diagnostics to active control.

We'll share results from a hospital pilot, including measurable energy reductions, smoother supply-air temperature control, and fewer nuisance alarms. Attendees will leave with a practical blueprint for building self-correcting reset logic in SkySpark, along with lessons learned on validating write-back actions in mission-critical, healthcare-grade environments.

BIO: Lincoln Harmer, PE, BCxP, is Vice President of Smart Buildings at kW Engineering, where he leads commissioning, energy management, and analytics-driven operations for building and infrastructure systems.

ABSTRACT: Fault detection tells us when equipment misses its setpoint. It does not tell us whether the building is meeting the requirements the owner actually specified, because those requirements live in the Owner's Project Requirements document, if they exist at all. The result is a persistent gap between what the rules measure and what the owner cares about: thresholds are hand-entered by whoever wrote the rule, drift is flagged without reference to design intent, and the question of whether a system was ever right at acceptance goes unanswered.

The presentation walks through this end to end and closes on the implication for how we develop Owner's Project Requirements and leverage them on projects. Attendees will leave with a concrete pattern for turning the OPR into the standard the analytics run against, and a clear view of where AI belongs in that architecture.This session argues that the gap is a data model problem, not an AI problem, and that the fix is to make design intent a first-class part of the model. Using SkySpark's Woolly assistant and its document ingestion capabilities, we show that a user can teach the platform an unfamiliar document type through a guided interview that yields Xeto specs, without hand-writing one. Each requirement then becomes a typed record bound to the equipment it governs. Once requirements are data on the equipment, the analytics already running on the same points can be evaluated against them directly, producing a compliance view against design intent rather than against arbitrary thresholds.

The presentation walks through this end to end on a live site, and closes on the implication for how we build rules: when the requirement is a typed record on the equipment, the rule reads it rather than hardcodes it. Attendees will leave with a concrete pattern for turning the OPR into the standard the analytics run against, and a clear view of where AI belongs in that architecture.

We think Cx deserves an independent data model, open and vendor-neutral, expressed in Xeto alongside the same ontology the community already uses for equipment and points. This data model allows the Cx record to live along side the building during operations. kW published one as open source and is building a Cx application layer on top of SkySpark with it.

The Cx process itself does not change. Checklists, functional tests, and issue logs work the way they always have. What changes is where the work lives: structured records linked to the assets, points, and documents they concern, instead of a binder that gets handed over and shelved.

We will walk through the open ontology and demonstrate where AI should and shouldn't be leveraged to augment the Cx process.

BIO: Aaron Hastings has more than 20 years experience in the Mechanical and Building Automation space. Consulting is a way to bring that expertise to make the rubber meet the road in the arena of large scale BAS implementation and building optimization. Recent speaking engagements include NexusCon 2025, 2025 fall ASHRAE BACnet Plenary group session for Network Security and Information Technology, Niagara Summit 2026, The US Department of Energy Better Buildings Better Plants 2026, Optigo Networks podcast 'How to Handshake' Episode 1, NexusCon 2026, and several others.

ABSTRACT: This session examines the financial and operational lessons learned from a machine learning system installed to optimize a university chiller plant. Lessons learned include overcoming financial, mechanical, and other. With a longer operating history now available, the focus shifts to evaluating payback and understanding the challenges of translating efficiency gains into verified dollar savings. The university’s main electrical feed, backup feed, and onsite solar complicate the calculation of electricity costs and avoided purchases. An effort to quantify savings through a vendor’s SkySpark integration revealed, after six to seven months, that the approach could not provide the required measurement accuracy. A move was made to a machine learning platform to visualize usage to spot billing errors, Use machine learning to backfill missing data, and use of machine learning to benchmark and predict future performance—enhancing decision making. If the notion of having AI in your building makes you nervous, this session will illuminate the ways this can be done safely—today.

BIO: Florin-Adrian Hebean is a Buildings & Energy Data Consultant with extensive experience in building automation, data engineering, and energy analytics. His career has evolved from BMS engineering to connected buildings and, subsequently, to developing data-driven solutions for building and energy performance.

Florin has been working with SkySpark since 2016 and has extensive hands-on experience with SkySpark and Axon programming. Over the years, he has contributed to hundreds of data projects and completed multiple SkyFoundry training and certification programs, covering SkySpark, Axon programming, and ViewBuilder.

He holds a PhD focused on the numerical modeling of an artificial intelligence-based controller for BMS applications, as well as Bachelor's and Master's degrees in engineering from the Technical University of Civil Engineering of Bucharest.

Florin has participated in several SkyFoundry conferences and SkyPosium events in Europe and the United States, sharing a strong interest in transforming building and energy data into practical, actionable insights.

ABSTRACT: Net-zero and highly energy-efficient homes combine multiple energy and building systems—including solar PV, heat pumps, smart metering, ventilation, underfloor heating, and indoor climate sensors—each generating operational data. The challenge is not simply collecting this data, but bringing it together, giving it context, ensuring its quality, and determining whether the home is performing as intended.

This presentation showcases a practical analytics approach for Net-Zero Homes using SkySpark as the analytics layer. It follows the journey from heterogeneous operational data to trusted and actionable information: data collection and contextualization, automated data-quality validation and cleaning, comparison of measured versus expected performance, KPI calculation, and automated Fault Detection & Diagnostics using SkySpark Sparks. Practical examples demonstrate how KPI Rules continuously translate time-series data into comparable performance indicators, while Sparks identify when and where performance issues occur—from low PV production and heat-pump performance issues to metering and communication failures.

The presentation then shows how these analytics support the complete process from measure, validate, and calculate to assess, report, and act. By transforming operational data into measurable evidence, SkySpark analytics can support energy performance assessment, reporting, investigation of deviations, and corrective action across a portfolio of Net-Zero Homes. The result is a scalable approach in which building data moves beyond visualization to become an operational asset for continuous performance management.

This presentation showcases a practical approach based on Net-Zero houses projects, using SkySpark to integrate data from multiple sources and turn it into actionable analytics. It covers the complete journey from automated data acquisition and Automatic Fault Detection to KPI calculation, dashboards, alerts, reporting and data visualization. A key focus is the role of the KPIs developed in SkySpark and the associated dashboards as the primary source of data and analytical results for the automated generation of EPF (Energy Performance Fee) reporting. The Net-Zero houses reporting application uses these results to provide the data required for EPF reporting.

The presentation demonstrates how energy analytics can bridge the gap between technical monitoring and real-world business requirements. By automating the identification of performance issues, calculation of relevant KPIs and generation of reporting outputs, large portfolios of sustainable homes can be monitored and managed more efficiently. The result is a scalable approach in which data moves beyond visualization to become an operational asset supporting both energy performance management and the EPF process.

BIO: Rick Jennings, PE, CEM, is an E-Mobility Product Manager and Data Analytics Application Engineer at SkyFoundry focused on applying the highly successful SkySpark platform as an analytics-driven energy management system for EV charging applications.

Rick has 12+ years of experience involving nuclear propulsion systems, railway electric traction systems, data centers, electric vehicle charging infrastructure, energy management, advanced data analytics, and IoT technologies.

In his previous role, he was ABB E-Mobility’s Product Marketing Manager and subject matter expert responsible for Terra 50 to 360 kW EV chargers in North America.

ABSTRACT: The Project Haystack ontology has seen tremendous advancement since last year as the working group continues to hone complex modeling concepts into standardized Xeto specs. Learn how Xeto features like mixins, globals, choices, and sugar specs have made their way into the Haystack 5 ontology. We'll also tour how the vocabulary has expanded to include new points for steam, refrigerant, doors, windows, and more.

BIO: Richard McElhinney is the Vice President of Technology at Conserve It, developers of PlantPRO™, a central plant control and optimisation platform, and international product developers and distributors for IoT, Building Automation, Smart Services and Big Data Analytics. Richard has over 25 years experience in product and solution development having worked globally with leading companies in the Smart Building Services space. Previously the Chief Software Architect at Conserve It and the Software Development Manager for Airmaster Australia, Richard has held numerous roles where he has delivered new and innovative solutions to the market. As a member of the open source Project Haystack community since its inception and subsequently as a Director of Project-Haystack.org, an open source, 501C trade association focusing on the standardisation and technology around semantic modelling and making equipment and devices self describing, Richard has contributed to numerous aspects of the development of Project Haystack. Richard continues to lead Conserve It to be at the forefront of central plant optimisation using Machine Learning and advanced mathematical modelling as well as furthering the goals and technology of Project Haystack to enable building owners and Facility Managers to unlock the value of data.

ABSTRACT: Haystack has become the common language for describing building systems, but creating a semantic model is still mostly manual work. Integrators tag points by hand at the edge, and then the same model is often rebuilt, remapped or re-tagged when the data moves to a cloud platform. Each step adds labour cost, delays projects and introduces inconsistency, which undercuts the interoperability Haystack was designed to provide.

At Conserve It, we took a different approach with PlantPRO, our platform for controlling and optimising central chiller plants. When a PlantPRO system is configured, it automatically generates a complete Haystack semantic model of the plant: equipment, points and relationships. Nobody has to tag anything. What the comissioning engineers already do to commission the plant becomes the semantic model, so it is accurate from day one and stays in step with the system it describes. This removes the cost of manual modelling at the edge and gives every site a consistent, standards-based model.

Creating the model is only half the story. In this session we'll show how PlantPRO connects to SkySpark in the cloud through our Sparkplug driver, using MQTT to publish both the Haystack semantic model and live data. SkySpark receives data that already carries its meaning, so the model never has to be built a second time. This has three main benefits. SkySpark deployments can scale across many sites without a matching increase in engineering effort. Customers save money by cutting out duplicated modelling work. And semantic models are more accurate because they come directly from the system that controls the plant, so nobody re-keys them by hand. Attendees will see how combining Haystack, Sparkplug and SkySpark creates a model-once, use-everywhere pipeline from the plant room to the enterprise.

BIO: Over 13 years in the controls industry, focusing on control systems development and integration, ranging from FDD, custom monitoring solutions, and advanced supervisory control. Focused on guiding clients to understand and deploy emerging technologies in different built environments.

ABSTRACT: Detailing the use of Phable for testing code implementation in SkySpark and as a middleware for other use cases.

BIO: Delaney Moore is a Software Developer II at ENFRA with three years of experience developing and supporting SkySpark-based solutions for building analytics, automation, and operational efficiency. Since joining ENFRA in 2023, she has focused on software development, platform improvements, troubleshooting, and scalable automation tools that help teams deliver high-quality SkySpark projects. Delaney works across SkySpark, Axon, and Fantom with a particular interest in improving workflows through automation and AI-assisted development. She is passionate about building maintainable solutions that simplify complex processes and help teams deliver greater value to clients.

ABSTRACT: Onboarding a new site into SkySpark is one of the most time-consuming and error-prone steps in any deployment — teams manually translate raw source documents into build files, then hope nothing was missed or mistyped along the way. This presentation walks through a working, AI-assisted pipeline that closes that gap end to end, turning raw documentation into a validated, build-ready site.

BIO: Manuel Morales is a Building Analytics Developer at Group14 Engineering, specializing in building analytics, monitoring-based commissioning, and advanced controls. He develops and implements SkySpark-based solutions that help building owners and operators identify performance issues, optimize HVAC systems and controls, and reduce energy use. His experience spans higher education campuses, national laboratories, and airport facilities, where he has applied analytics and supervisory control strategies to solve complex operational challenges. Manuel brings a practical, hands-on perspective to using building data and SkySpark to improve system performance and advance high-performance building operations.

ABSTRACT: Lawrence Berkeley National Laboratory (LBNL) and Group14 Engineering (Group14) will present together on LBNL’s supervisory control research program and Group14’s implementation of ASHRAE Guideline 36 (G36) air handler sequences using SkySpark.

Since 2019, LBNL has partnered with leading providers of smart building analytics to pilot fault-free optimal control (FFOC) in real building conditions. FFOC refers to a suite of capabilities to streamline and automate the correction of common HVAC faults, and the implementation of advanced HVAC control strategies. These capabilities are implemented via SkySpark among other analytics software offerings tied to a building's Building Automation System (BAS). The spectrum of strategies ranges from one-time fixes, to recurrent active testing, to continuous supervisory updates of control parameters such as setpoints used in air handler unit (AHU) trim and respond sequences found in ASHRAE G36.

Group14’s client, the University of Colorado, wanted to adopt these G36 reset strategies throughout campus, however traditional full-scale BAS reprogramming proved prohibitive due to the high costs and limited contractor availability. Leveraging CU’s existing SkySpark project, Group14 utilized LBNL’s existing Axon codebase to implement AHU supply air temperature and pressure reset strategies in SkySpark for two existing building air handlers on the Anschutz and Denver campuses. This presentation will discuss CU’s motivations for implementing G36 sequences, Group14’s deployment of these sequences, energy savings achieved, and the lessons each partner learned from this project.

BIO: Rob Patterson is a Business Development Manager at Miller Electric Company. He retired from GE and ABB after 35 years, during which he worked as a developer and contributor to monitoring solutions including GE Envisage, the Proficy version of Trane Tracer, and ABB Zenon. He continues that work today with Miller PRISM. Rob lives in Southwest Pennsylvania and enjoys hunting, fishing, and farming.

ABSTRACT: Modern facilities generate large volumes of operational data, yet that information is often isolated across electrical systems, HVAC equipment, generators, transfer switches, PLCs, meters, and other intelligent devices. This presentation explores how Miller Electric Company developed PRISM as a vendor-agnostic facility monitoring and asset-management solution built to unify these traditionally disconnected systems.

Using an open development model and SkySpark-based analytics, PRISM connects diverse facility assets through an on-premises gateway while providing secure web access, key performance indicators, alarm notifications, and machine-learning capabilities. The platform combines real-time monitoring with contextual asset information, including maintenance history, equipment documentation, electrical one-line diagrams, labels, nameplates, BIM models, and power-system study data.

The session will discuss the practical questions organizations should address when implementing facility monitoring: who is watching the system, whether they know what to look for, how cybersecurity and network independence are maintained, and whether the solution can scale across equipment manufacturers and facility types. Attendees will learn how PRISM supports asset-health visibility, energy-cost monitoring, maintenance and work-order tracking, automated notifications, optional remote control, and continuous monitoring through Miller’s Client Services Center.

The presentation offers a practitioner’s perspective on transforming raw building and electrical data into actionable operational intelligence while preserving flexibility, interoperability, and customer ownership of the solution.

BIO: Ross Schwalm is an IoT and Data Enthusiast at SkyFoundry. Ross has over 20 years of experience focusing on IoT, cybersecurity, and mobile technologies while working in different roles, including consulting, sales engineering, business development, product management, and policy support. At SkyFoundry he builds new features as a developer, teaches people how to use SkySpark as a trainer, solves problems in tech support and highlights SkySpark's benefits through sales support.

ABSTRACT: SkySpark is adding a plethora of new security features this year. Attend this session to learn about upcoming OAuth 2 support and new features to streamline key rotation.

BIO: Leroy Simms brings a comprehensive, end-to-end perspective to building intelligence—combining 25+ years of experience across mechanical diagnostic service, automation engineering, energy consulting, and advanced data analytics. As Co-Founder, Analyst, and Consultant at ABI Advantage, Leroy focuses on custom SkySpark Axon programming, multi-protocol system integrations, and actionable FDD workflows. Holding a B.S. in HVAC Engineering alongside Master licenses in HVAC and Electrical, Leroy bridges the gap between deep technical field domain knowledge and cutting-edge data analytics to transform raw building operational data into sustainable operational efficiency.

ABSTRACT: As state and local Building Energy Performance Standards (BEPS) mandates tighten across the country, building operators face a complex maze of conflicting regional energy metrics and steep financial penalties. This session demystifies current BEPS programs and outlines practical methods to continuously model, monitor, and report compliance status using SkySpark. Attendees will learn how to turn fragmented energy data into an automated, proactive compliance strategy that safeguards against operational and financial risk.

Key Takeaways:

• *Navigating Regulations: Understand core nuances and metric variations across emerging regional BEPS programs.
• *Continuous Monitoring: Implement SkySpark workflows to track performance against compliance baselines in real time.
• *Penalty Avoidance: Establish automated reporting to catch non-compliance early and prevent costly fines.