Pallas Analytics

Understanding agency and control over innovation

Pallas Analytics is a working archive of prototypes and research on the future of apprenticeship-based work in the early years of technical careers. Our work explores how critical thinking, systems design, and architecture interplay with new and evolving technologies. Traditionally, our prototypes focused on the data. Lately, we are leveraging agentic AI in creating working systems as quickly as possible. In that process, we uncover the gaps and requirements to transform an agentic idea into a functional, differentiated product. Our thesis remains the same: good products and sustainable technology will always rely on clearly defined designs, an understanding of data sources, and complete control over system architecture. The projects here are experiments in making complex decisions more visible: regulated model review, GxP execution, publishing intelligence, climate-to-sensory risk, and the future of work.

Prototype archive

Projects

These projects were commissioned to explore feasibility. Our approach is to understand the problem, define the architecture, source the data, and then build prototype applications using agentic AI, since we have found that we can quickly build in a matter of days, if not hours. After building, our core work is to identify the gap between prototype and product from a technology perspective.

01 regulated operations Regulated Pipeline CRM This project was built for a biotech startup with limited resources on quality control personnel. We transformed validation plans, requirements, approvals, and operating signals into a live GxP execution pipeline. The end result became a "Control Center" for product and process owners to coordinate and manage the development pipeline of a GxP regulated product. 02 regulated AI Regulated AI Sandbox Within RFP-driven sectors, we wanted to accelerate market entry for smaller machine learning companies to compete for contracts. We designed a governed model evaluation environment built for reviewability, evidence, and deployment decisions. In this system, data sharing is managed based on security clearance, and clients are able to continuously provide model updates in a competitive "Kaggle-style" setup. We liked this idea in helping industry quickly identify promising and state-of-the-art work rather than defaulting to the "old guard." 03 publishing intelligence Translation Scout Moves foreign-market coverage and rights signals into a ranked scouting pipeline. 04 critic intelligence Good Reviews Ranks reviewer and critic fit from industry taste data, review data, source catalogs, and taste vectors. 05 climate and sensory data ClimateSensory Explores climate, sensory quality, and supply-chain risk for wine, coffee, vanilla, vetiver, lavender, and perfume materials.

Future of work

The Future of Career Development in STEM

Lately, our research focus has been on understanding how entry-level work and the nature of apprenticeship in junior STEM roles have changed. Our view is to transform this into curriculum innovation at university curricula.

01 working paper The Junior Analyst Problem What changes when entry-level software engineering opportunities are automated or abstracted away. 02 working paper The Hidden Curriculum of Office Work The informal skills students learn only after entering professional environments. 03 working paper The AI Apprenticeship Audit A framework for where AI supports, weakens, or reshapes learning on the job. 04 working paper From Prompt Literacy to Professional Agency Moving beyond tool fluency toward judgment, authorship, and critical adoption.

Method

What deserves to become a product?

In-depth discovery on processes with clear automation potential.
Prototyping with a design-centered workflow and clear, visible pipelines where the user understands the product implementation.
Trace the data sources, assumptions, and gaps instead of hiding them behind a polished demo.
Let the prototype reveal what architecture, governance, and product surface the domain actually needs.

Contact

Prototype development, research, and applied AI systems.

Reach out to discuss prototype development and research collaborations on the future of work. We are happy to share our approaches to the work highlighted above and further research in any of these areas.

Open the contact form