These studies empower my consultations and technical sales. Select from the links below to view research and results.
- Select a topic or scroll through them all -
The Need for Human Expertise at the Dawn of AI Code Generation
Researched multiple approaches to writing code via Generative AI. Led an internal study with 25 engineers, testing twenty AI solutions across the SDLC. Gathered results over thirty days and made recommendations based on engineering insights. Culminated into a thought leadership article.• Created multiple apps in AI myself to prove the capacity of the approach wile providing novel solutions and then training other engineers to do the same.• Researched solutions for customer safety and documented a best-practice approach with 7 security gateways.• Produced a video that demonstrates the possibilities, shared during a summit of 77 agencies, positioning our as a leading AI-first enterprise.Essential Findings: Human expertise and oversight are critical during the infancy of Generative AI Coding. New specialized skills are needed to manage the AI paradigm effectively.
An app generated with AI, featuring sophisticated UI / UX.
3 day build.
Native, React Native, and a Novel Solution to Bridge the Gap
Wrote a white paper on the differences of Native and React Native capabilities and outcomes for an enterprise Fintech client in mid 2022. This paper changed the company's non-informed strategy and empowered them to adopt a novel React Native architecture that would not interfere with existing iOS and Android code bases. Sourced a React Native expert to build the designed solution, saving the client >$250k in migration costs, delays, and potential ongoing complications.This white paper became widely referred to and was used in client discussions and pre-sales conversations. Was planned to be used as a keynote address to CTO's at a summit in NYC.Essential findings: React Native provides an excellent path to cost reductions via a single code base, yet one must color within the creative limits of the RN system or face costs and timelines even greater than multi-platform native development.
Slides from a presentation given to the fintech client regarding Native vs React Native.
Generative AI Video Exploration
Led research for a 2K person agency in the use of AI across the SDLC, including usage in more creative outputs such as imagery and video.• Created a 1-minute promo video entirely with AI generated content as a first time user of Generative AI Video.• Gave myself a single day to see what could be accomplished in 24hrs when starting with no prior use. Captured the learnings in an informative video for the agency, to spread the insights gained.• View the video below.
Created a 1 minute promo video with Generative AI. Made a tutorial video to explain how it was done.
1 day build.
Monorepo or Multirepo?
Researched a monorepo migration solution for an enterprise sport client who adopted the plan.• Leveraged their existing knowledge base and combined it with a 3 month dive into the subject matter.• Studied the full history of "repository as a service", noting trends and drivers.• Explored the strengths and weaknesses of both the Monorepo and Multirepo approaches, as well current hyper-modularization trends and outcomes.• Provided clear benchmarks based on a comprehensive proof-of-concept that elevated the research from theory into fact.• Suggested a course of action based on client needs. The client decided to move forward with the new strategy.• Provided exacting steps for the migration, including how to set up relevant security protocols, and how to migrate gitActions.• The resulting documentation, as well as the demo, were lauded as the best they'd even seen, internally or externally.
A presentation detailing how to transition from Multirepo to Monorepo.
Unlocking SCIENCE: The Key to Transforming Your Business with AWS Cloud
In today's rapidly evolving digital landscape, businesses are constantly seeking innovative solutions to stay ahead of the competition and meet the ever-growing demands of technology and data management. The journey towards digital transformation can be complex, but the path becomes clearer when we explore the AWS Cloud through the lens of "SCIENCE". This acronym encapsulates the core reasons why moving your business data to the AWS Cloud is not just a strategic move—it's a transformative one. Let's delve into the essence of SCIENCE and discover how it unlocks a world of possibilities for businesses aiming for growth, efficiency, and scalability.
AWS Cloud prioritizes security above all, offering a robust infrastructure designed to protect your data, applications, and resources with multiple layers of defense. Additionally, AWS ensures global compliance with a wide range of international and industry-specific standards, providing businesses with the assurance that their data handling practices meet stringent regulatory requirements worldwide.
AWS Cloud revolutionizes cost-efficiency by leveraging economies of scale. With a pay-as-you-go pricing model, AWS enables businesses to switch from capital expenditure (CapEx) to operational expenditure (OpEx), ensuring that you only pay for the resources you use. This flexible approach to cost management empowers businesses to scale and adapt their operations with financial agility.
AWS Cloud is synonymous with innovation, offering hundreds of powerful features and services that are immediately available to new companies. This extensive suite of capabilities covers everything from computing power and storage to artificial intelligence and machine learning. AWS empowers companies to stay ahead in their respective industries by leveraging the latest in cloud technology.
AWS Cloud exemplifies next-generation efficiency, with a finely tuned ecosystem designed to eliminate unnecessary costs. Advanced monitoring and management tools allow for real-time adjustments. Through the Management Console, businesses can swiftly modify and scale resources. This level of control and efficiency allows companies to focus more on innovation and growth.
AWS's network excels in High Availability and Reliability, featuring Auto Scaling and Fault Tolerance to ensure seamless operations and handle traffic spikes without downtime. AWS's Globally distributed data centers enable businesses to position their applications geographically close to their customers, reducing latency and meeting regional compliance needs efficiently.
Customization is the heart of all AWS services, offering unparalleled control over data and infrastructure. Detailed tracking and analytics tools enable organizations to monitor performance, usage, and costs in real time. This level of insight supports continual improvement, allowing companies to optimize, innovate, and deliver enhanced services to their customers.
Elasticity, a cornerstone of AWS's Elastic Compute Cloud (EC2), not only allows for the dynamic scaling of computing resources to match user demands but also plays a critical role in disaster recovery and business continuity. Resources can be quickly adjusted or relocated across different availability zones, facilitating rapid recovery from disasters, making it an ideal platform for prioritizing preparedness.
Twelve Legacy Code Troubles and their Six Solutions
Proposed the modernization of an enterprise sport client's legacy code base. Won the work and spent a number of weeks running interviews with stakeholders and engineers, reading existing documentation, and exploring the code. The result was a 75 page white paper that detailed their 12 signature problems and proposed 6 solution goals. The client decided to move forward with elements of the strategy.HISTORICAL PROBLEMS
1. Big Ball of Mud code base, a design anti-pattern that lacks perceivable structure
2. Pop-Culture Code, replete with artifacts of long lost engineering trends
3. Conflicting Priorities among leaders and teams across 33 projects and tasks
4. Conflicting Strategies that fight for monorepo or multirepo methodologies
5. No Single Source of Truth - No central repo for management metrics and info
6. Dearth of Metrics - essentially no visibility into critical processes to inform decisionsTECHNICAL PROBLEMS
7. Code Placement - poor modularity resulted in poor code organization across packages
8. Slow Builds - Due to hyper-modularization, package resolution is slow and fragile
9. Package non-resolution - builds can be so slow that they sometimes fail to resolve at all
10. Standardization - Universal adherence to linting and automation standards needed
11. Availability - The app is hovering at 200mb, the limit for full cellular availability
12. Instability - Unit test coverage is as low as 10% and visibility metrics are not active6 SOLUTION GOALS
1. Observability - Must be active, targeting clean metrics, and available in a dashboard
2. Uniformity - Modernizing languages, patterns, and linting to boost performance
3. Dependency - Rework repository structure to minimize long build times and fragility
4. Productivity - adopting reusable components to minimize defects development time
5. Reliability - Enforce Unit Tests, increasing coverage and protecting modernizations
6. Clarity - Establish a central knowledge base and enforce documentation for all featuresThe white paper clarified all of the above bullets, both problems and solutions, and provided success criteria for each.The paper provided 2 strategies to complete the modernizations in a reasonable time with a right-sized team.The paper also included other helps and deep dives into:• Swift Package Management
• The Compile Processes
• Clean Layered Archicture
• A wide array of dependency solutions
• Case studies (especially regarding monorepo adoption)
• Principles of development (SOLID, etc.)
• Metrics of developments (Cyclomatic Complexity, etc.)
• Codebase Audit Strategy (What to look for and capture for modernization)Regarded as the best white paper the team had ever seen.
A presentation detailing the 12 signature problems of the code base, and 6 goals to solution them.
LMS - A Crowded Field Meets a Novel Use Case
Won a contract with Sutter Health to build a PWA Mental Health and Wellness app for teens and young adults. Consulted their product team regarding many third party providers across a spectrum of services, with LMS being chiefest among them.There were 2 unique needs that had to be met. The first was the ability to access and manage the LMS by API only, allowing our designers to own the visual experience. The second was a novel security approach I designed to protect PII to the highest degree.• Researched 105 LMS providers based on capacity, platform age, customization, security, reliability, cost, and existing clientele. Identified the top 3 and interviewed their sales and tech teams via zoom.• Selected the top LMS and negotiated terms on behalf of the client, saving them significant funds. Handed off the LMS to the client for final negotiations, security reviews, and onboarding.• Negotiated a special window of time in which our engineers could work within the LMS while final negotiations and security reviews were happening, saving months of delivery delays.• Consulted between the client, our engineers, and the LMS throughout the development process to maintain high standards of security, bandwidth speed, api reliability, and end user experience.
A presentation detailing 105 LMS' studied, and the selection criteria for the winner.
A Shared Component Strategy for Multi-Disciplinary Teams
Proposed and then led the creation of a shared UI component library for an enterprise Fintech client, mirrored across iOS and Android. Through the build and implementation phases spent a good deal of effort studying best practice. Gained critical insights that led to the overwhelming success of the initiative, with 40 components and 400 variants built over six months by an interdisciplinary team. Through use of the library we achieved a 57% reduction in feature development time.• Shared component ideation and construction requires an intricately woven flow between disciplines. Researched and created an interdisciplinary workflow with clean handoffs and clear delineations of roles and responsibilities, providing an efficient pipeline from Figma to delivery.• Added a new VQA process in which Design visually vets the builds before they are delivered. This is in addition to standard QA. The client sought pixel perfection, and VQA achieved it, reducing visual defects by 90% and mitigating defect loops.• Researched and created a preview tray that can be accessed from anywhere within the app during debug. All shared components were listed in the tray and could be launched to the screen, with additional controls to affect tokens, icons, line counts, fonts, colors, and etc. This empowered the team to easily test and experiment with live components without needing an expensive dev/test/adjust loop, and without needing to search the app for a natural example of a component in the wild. The preview tray reduced VQA times by 75%, and greatly empowered development efforts.• This knowledge has gone on to service other clients and reduce their development times.
A presentation detailing 105 LMS' studied, and the selection criteria for the winner.
From High-Level Code to Machine Execution: Navigating the Compilation Steps
Researched the compile time process for iOS applications to share with a client the critical impact of poor code on build times. Created the infographic below. This was coupled with a white paper that showed what both good and bad code look like at each stage of the process, to better understand the CPU burn that happens as bad code bloats through the 10 steps.• Included examples of code smells that lead to bloat and how to solve them.• Was one element of an overarching optimization that can speed up build times by 50%.
Improved Sprint Sequencing for Maximum Velocity
• Researched and designed an improved sprint mechanic to allow for more stable workflows. A continual problem in standard Design / Development sprint offsets is the constant tardiness of designs, which eat into development time and cause delivery delays and engineering downtime.• In more typical fashion Design has sections of a user experience to finish per sprint, after which they must lock down the design and allow Product to write the User Stories (pictured below). Engineers would pick up these designs and stories at the beginning of their sprint, but often ended up waiting for designs and stories for days on end.• Researched and added a new Product sprint offset (pictured below). This created a buffer between development and design during which Product could wrap up User Stories. Decreased downtime by %25 and increased delivery accuracy by 40%.• Saved considerable time and money for the client, and went on to be used in every project I led going forward.
A presentation explaining a new Product sprint offset to reduce friction between Design and Dev.
The Limits of Asyncronicity, and the Criticality of Connection
Researched cross-timezone workflows on behalf of an enterprise Fintech client. The engineering teams continually experienced 24 hour feedback loops (read: delays), where a team in one part of the world paused work for an entire day due to unclear requirements and had to wait until the next day to get answers to their questions.• We first spent time nailing down Jira tickets with a concept called The Definition of Done in which a ticket was not considered ready for work until it contained ALL relevant info. This helped mitigate some of the feedback loop, but the loop persisted.• There was no silver bullet to fix this. No asynchronous process completely staunched these time bleeds. The problem is essentially human, not merely process, and questions always remained, despite how verbose the instructions became.• The answer was real-time human connection.• For sensitive or complex work, handoffs had to be in realtime to allow for conversations and for questions to be fully answered. These handoffs were chiefly handled by engineering leadership, protecting the wider teams from the time stretch.• U.S. leadership would meet with India teams in the late evening to confirm requirements. And India leadership would meet with U.S. teams the following morning to report progress and blockers, keeping the feedback loop as tight as possible.• Cross-timezone training sessions preceding the development of each feature further paved the way for clean handoffs.• As human connections grew across timezones, the need for real-time connections diminished.• While clear documentation of requirements and success criteria are mission critical, human connection and conversation is the single biggest preemptor of costly feedback loops.
A presentation explaining best practice for overcoming the 24 hour feedback cycle with globally distant teams.
Managing the Transition from Agency to Staff Augmentation
Transitioning an agency's in-house team of engineers to an external staff-augmentation model, especially when the client remains the same, requires careful planning and execution to ensure continuity, maintain morale, and achieve seamless operational integration. Here are the steps researched and proposed to the client to manage the transition effectively.• Conducted a detailed assessment of the in-house team's structure, roles, responsibilities, and the specific needs of the client. Developed a phased transition plan that addressed the scope of work, timelines, and other specific client requirements.• Communicated clearly about the reasons for the transition, the benefits, and how it will be managed. Addressed concerns and provided clear information about changes in reporting structures, roles, and expectations.• Defined roles and responsibilities of the transitioning team members within the new staff-augmentation model.• Integrated tools and systems, ensuring the team had access and any necessary training.• Fostered team integration through regular communication channels to build trust and collaboration between the client team and the augmented staff.• Regularly assessed the performance of the staff-augmentation model against predefined metrics. Made adjustments as needed.• Hired for additional roles, and moved unwilling talent from the team roster to other projects or the bench.• Provided ongoing support, training, and professional development opportunities to the augmented staff. This ensured they remained motivated, engaged, and aligned with the client's long-term goals.• Periodically reviewed the staff-augmentation arrangement with the client to evaluate its effectiveness and identify areas for improvement. Adjusted team composition, roles, or processes to better serve the client's objectives.
Minimize Visual Defects and Attain Pixel Perfection
Pioneered an enhancement in the software development lifecycle for a leading enterprise fintech client aimed at achieving unparalleled pixel perfection in final software builds.• Conducted extensive research into the root causes of visual defects slipping through the existing quality assurance processes, identifying a significant gap in the detection of design-related issues that resulted in technical debt and recurring defect loops.• Leveraged findings from my research to propose and advocate for the introduction of a Visual Quality Assurance (VQA) stage, a concept designed to specifically address and mitigate the oversight of visual defects.• Spearheaded the design and integration of the VQA process, ensuring it was strategically positioned after the development phase but before final delivery.• This initiative was grounded in the insight that Designers, with their deep understanding of both the aesthetic and functional intent of designs, were ideally placed to identify and rectify visual discrepancies.• Led cross-functional collaboration, engaging Product, Design, and Engineering teams in the adoption of the VQA step, which was a direct result of my initiative and research-driven approach.• Through the implementation of VQA, achieved a transformative 90% reduction in visual defects pre-delivery.
A presentation explaining a new process step called VQA which reduces visual defects and defect loops.
Why Web AR Is The New Ecommerce
Researched and wrote an article on the rising power of WebAR, with a strong marketing perspective.A summary:Shopping is undergoing a revolutionary transformation thanks to WebAR (web-based augmented reality), making it possible to try on shoes, browse celebrity wardrobes, and see art in your space with just a tap on your browser. This cutting-edge technology not only enhances online shopping by making brand messages twice as impactful but also significantly boosts sales and engagement. With successes from Saatchi Art, Shopify, and even giants like Amazon and Nike, the potential of WebAR is undeniable. Yet, despite its proven benefits and consumer demand, adoption among retailers is low. This article delves into how WebAR is reshaping shopping, backed by substantial investments and innovations, urging retailers to embrace this change or risk becoming obsolete.The article was picked up by Media Post and featured as the first piece in their new 3.0 Platform.Read the full article here.
AR Limitations of Non-Lidar Devices
Researched the efficacy of AR experiences on LiDAR vs non-LiDAR devices, along with creating a compelling PoC. Counseled an AR / AI company with the results. This was during the advent of LiDAR. Encouraged them to target LiDAR devices if they sought to offer world-class, breakthrough AR experiences, and warned them of the danger of over promising AR capabilities on non-LiDAR gear.A summary of the findings:Non-LiDAR Devices: Limitations in AR• Depth Perception: Struggle with accurately gauging distances and understanding the depth of real-world environments, leading to less immersive AR experiences.• Object Detection: Have limited capability in precisely identifying and interacting with physical objects, affecting the realism of AR overlays.• Spatial Awareness: Offer poorer spatial awareness, making it challenging to place virtual objects in real-world settings in a way that they appear naturally integrated.• Low-Light Performance: Perform poorly in low-light conditions, as they rely heavily on visual data that can be compromised without adequate lighting.• Accuracy and Resolution: Provide lower accuracy and resolution in mapping environments, which can result in virtual objects appearing out of place.LiDAR Devices: Enhancing AR Experiences• Advanced Depth Sensing: Use direct time-of-flight measurements for superior depth perception, allowing for more accurate and realistic placement of AR objects in the real world.• Improved Object Recognition: Better at detecting and understanding the geometry and surfaces of objects, enabling more interactive and engaging AR experiences.• Enhanced Spatial Awareness: Provide greater spatial awareness, crucial for complex AR applications like indoor navigation and augmented reality gaming.• Superior Low-Light Performance: LiDAR technology is effective in low-light conditions, broadening the environments in which AR can be used effectively.• High Accuracy and Detail: Offer higher accuracy and detail in environment mapping, supporting more seamless integration of virtual and real-world elements.• Future Potential: LiDAR is paving the way for next-generation AR applications, including more detailed environmental interaction, advanced gaming experiences, and practical uses in education, healthcare, and industry.
The Impacts of Augmented Reality in the Future of Fine Art
Extensively researched the rapidly scaling market trends of Augmented Reality in 2018.• Wrote and designed a 20 page investor document explaining the future of AR.• With this document, and some presentation work, secured a $300k investment for a purchase of 10% of my company Varlio.
The Augmented Arts debut of Varlio in Manhattan Beach, 2019.