Intersect Next

Fred Bliss: At the Intersection of Applied AI, Enterprise, and Technology

Offering personalized AI Guidance, Hands-on Engagements, Insights, and Research tailored to organizations from startups to enterprises.

Fred Bliss's expertise is shaped by decades of experience in complementary industries and technical fields, including over two decades in data consulting, as well as theco-founder and CTO of Aptitive, a modern data & analytics consulting firm founded in 2013.

As CTO and co-founder, Aptitive grew to a team of ~55, often doubling each year, with a successful exit in 2021.

If you or your organization have an 'AI mandate', are spinning up a new project, or need a second eye in delivering a tangible outcome, Fred can help you re-think how to look at various use cases - including identifying and separating what's not yet solvable and/or high risk - and trading it for more granular building blocks that will deliver outcomes and set a foundation for the future. Given the large number of great use cases out there, sometimes it's often a matter of perspective and nuance - and getting the right advice on how to go from Start to Finish.

AI Guidance, Advisory, Insights, and Research

Approach and Perspective

Fred has made it his passion to be on the bleeding edge while keeping one foot grounded in enterprise realities. This gives him a unique perspective into the future possibilities and their fit into different types of organizations. By continuing to be heavily hands-on, it becomes easier to see great ideas that work, and others that simply sound great on paper.

Experience and Network

Fred leverages his unique perspective, extensive cross-industry experience, and network within the emerging modern AI ecosystem to provide valuable insights and guidance.

Navigating the Frontiers of AI Potential

In this era of rapid AI advancement, organizations face unprecedented opportunities for innovation and efficiency. As your advisor, I can guide you through the landscape of possibilities, helping you identify and prioritize AI initiatives that align with your strategic goals:

• Synthetic Data: We're already seeing the frontier labs heavily using synthetic data (or, data augmented by other data via LLMs). The possibilities here are the most exciting ones that most organizations already have the talent and know-how to build. From transforming checklists or workflows, such as audit or compliance workflows, into code for quicker evaluation. Or making your product data shine by taking all that rich data on marketing brochures and sales training guides, or narration and imagery from product demonstration videos. From generating or inputting images, audio, or other unique and creative sources of data, and augmenting or creating new sources of data, the only limit is creativity. The possibilities here are endless and the most exciting to me, largely because there's few barriers to getting started immediately.
• Multimodal AI: With native multimodal (text, images, video, audio, etc) models now entering the market, as the capabilities grow, it will soon become the standard, allowing for 'any' input to 'any' output, be it an image or audio capture or video, or all the above together. Today, using some methods that were only recently uncovered on the frontier, we can use vision models in conjunction with advanced retrieval (search) models, allowing for the vision model to natively understand complex real-world documents. While direct 'question answering' has been part of the 'mishype', there's already immense value in being able to query vast amounts of rich documents, including slide decks and PDFs with visual elements (e.g. 'what was the YoY sales revenue chart we shared on investor decks' can now typically return exactly the right page you're looking for). From there - it's a matter of building creative new interfaces to interact with the highly relevant retrieved data in a brand new way. We can go way beyond chatbots and 'give me the answer' - particular given nuance and context. Understanding the current limitations and potential of these technologies is crucial for extracting valuable insights.
• Advanced Retrieval and Search Capabilities: Improve your organization's ability to access and utilize information effectively. Unless you're a technology or software company, building and maintaining a vector database and search engine (and the 'glue' between it all) from scratch can be a costly and unnecessary endeavor. Instead, let's explore more practical and efficient solutions that can meet your needs.
• AI Engineering as a new Role: The rise of the AI engineer as a formal role has been growing substantially over the years. With firsthand experience playing that role, identifying the right people on your team that can combine data engineering, software engineering, and AI engineering into a cohesive pipeline is an easy 'win' that can be accomplished today, and reap long-term benefits in return. The right approach here can lead to continual improvements in new and existing pipelines without unnecessary complexity.
• Business Process Automation Opportunities: Before anything else, have your team members map out their day-to-day processes and identify areas where automation can help simplify their workflow, eliminating tedious tasks and freeing up time for more strategic work. Once you build that muscle, you can move on to broader, more complex processes - identify key success metrics, and translate these into code, data, feedback for AI reward models, and you're now on the path toward aligning AI capabilities with your business objectives. As part of our advisory relationship, we can quickly show brand new (yet very familiar) user interfaces that give business users unprecedented control and exploration.
• Creating a 'Data Flywheel' : Develop a systematic approach to AI engineering that creates a self-sustaining cycle of improvement, where high-quality data, robust infrastructure, and reliable deployment processes work together to accelerate AI development, deployment, and iteration, ultimately driving business value and competitiveness.
• Measure, Learn, Adapt: Building Foundations for AI Excellence: You can't optimize what you can't measure - if you don't know what 'good' outputs look like from your AI workflow, it's difficult to optimize anything before that. However, by capturing inputs and outputs, we can jointly identify what 'good' looks like using real-world examples, creating a crucial feedback loop. Start by logging and capturing key data points, even if your definition of success is still evolving. This foundational work enables continuous improvement, allowing you to refine your AI initiatives and adjust course over time.
• Data Privacy-preserving AI Solutions - a Hybrid Approach: Explore on-device and edge AI implementations to maintain data privacy and enable low-latency applications crucial for sensitive operations. Combined with 'smart' routers, users within your organization can leverage on-device models for everyday common tasks, while offloading larger tasks to a wide variety of different models that can vary by use case, workflow, budget, and capability. Given the expectation that models will continue to commodify, with Compound Systems growing, having a vision and playbook for this ever-changing landscape is of considerable value.
• Interactive UIs in Minutes and Hours, Not Days: As we've seen in other tech cycles, clever and creative user interfaces will become the key drivers of AI adoption. Moving beyond chatbots and using LLMs in new ways requires new interfaces - and new interfaces require creativity. Explore opportunities to create bespoke, purpose-built AI-driven UIs that tell a story and showcase your ideas and vision. Business analysts with little-to-no experience in building frontend code can now provide the rich business context and data as the driver for AI-generated UIs, for everything from prototypes to ideation. With the ability to create custom UIs in minutes or hours, you can turn meetings into immersive, interactive demonstrations of your vision or idea.

It's important to note that these capabilities are achievable today with the right approach, expectations, and technical considerations - specifically taking advantage of what we know to be gaining real traction amongst trusted sources of research and projects in the field - while simultaneously aimiming for simplicity and practicality. Getting these nuances right is often the difference between an AI initiative's success and failure. Simpler, smarter solutions can often provide significant value without the complexity tradeoff - but it's important for you and your team to be aware of the tradeoffs, so that your organization can make informed decisions.

My role is to help you navigate these nuances, set realistic expectations, and determine the most effective approaches for now, the near term, and the future. The ideas presented here are just a starting point - I'm happy to explore more possibilities tailored to your specific needs in our conversations.

Now is the time to strategically integrate AI capabilities into your organization, but it must be done with a clear understanding of current possibilities and limitations. This approach represents an investment in your organization's future, grounded in practical, achievable steps.