Max Boving.

Personal portfolio / 2026

Portrait

Education

B.S. in Information Sciences
Emphasis in Data Science
Minor in Business Administration

About

About Max Boving

I’m an AI and data engineer who turns ambiguous product and data problems into working systems across retrieval, high-volume data platforms, applied modeling, APIs, and full-stack products.

My work spans production RAG, GraphRAG prototyping, vector search, retrieval evaluation, prompt testing, and LLM cost and latency analysis. I build with Python, SQL, TypeScript, FastAPI, React, PostgreSQL, Redis, Docker, AWS, Azure Data Lake, Microsoft Fabric, and Cloudflare.

Recently, I diagnosed retrieval failures in an AI customer-support platform, rebuilt a jury-persona system around Census and survey data, migrated more than 80 million environmental sensor records, and built tariff intelligence pipelines for supplier decision-making.

I’m most interested in forward-deployed engineering, data science, and applied AI roles where I can combine technical ownership with product thinking, client collaboration, and real-world problem solving.

References

A few words from people I’ve worked with.

“Max is one of those rare people whose drive, depth, and vision stand out well beyond his years. I’ve had the privilege of working with him over the past two years as an intern, student worker, and now as a company co-founder, and he consistently brings both the ambition to create and the determination to see projects through to completion. I’m excited to continue building alongside him and see what he accomplishes next.”
Michael WagenheimAssistant Dean of Information Technology & Online Learning
James E. Rogers College of LawLinkedIn ↗
SaticoyFull Stack Engineer (Contract)AI Customer Support Platform

Saticoy is an AI customer-support platform.

  • Diagnosed retrieval failures caused by oversized and irrelevant context chunks, which increased token usage and contributed to inaccurate or hallucinated responses.
  • Built reusable prompt libraries and added GitHub Actions CI tests to validate context formatting and structure before changes reached production.
  • Prototyped a GraphRAG workflow on customer-shaped fixture data, improving the latency and cost profile of multi-hop queries compared with the existing approach.
OptoJuryCo-Founder & Technical LeadJury Simulation & Trial Intelligence

OptoJury is a jury simulation and trial intelligence platform.

  • Founded and built OptoJury from the ground up, owning persona development, simulation workflows, the full-stack application, and cloud infrastructure.
  • Rebuilt the persona system after attorneys raised concerns about AI-generated jurors, grounding profiles in Census demographics and established survey data instead of invented biographies.
  • Built a statistically checked pipeline that adjusts population records to county demographics, identifies recurring attitude groups, and generates locally representative panels now under early attorney review.
HighBeam AnalyticsData Analyst & AI Engineer InternTariff Intelligence

HighBeam Analytics builds tariff intelligence tools for trade and sourcing decisions.

  • Built a searchable prototype tariff tracker and scheduled pipeline that joined product, SKU, and HTS data in Azure Data Lake Gen2 and Microsoft Fabric.
  • Combined LLM classification with regex rules and defined heuristics to suggest likely HTS matches and flag uncertain classifications.
  • Modeled future landed costs with Monte Carlo simulations and presented tariff exposure in Power BI with conditional email alerts.
Biosphere 3Technical LeadEnvironmental Data & AI Platform · University of Arizona

Biosphere 3 is an environmental data and AI platform.

  • Turned an uncertain research-platform idea into a near-real-time scientific dashboard, migrating more than 80 million sensor records from a fragile legacy system into PostgreSQL.
  • Built materialized views with 15-minute refresh jobs and designed the API layer, reducing key database queries to approximately 1 ms across more than 10,000 sensor streams.
  • Developed a RAG prototype combining sensor-table queries with retrieval from embedded scientific documents, and led five contributors through task assignment, code review, architecture decisions, and stakeholder coordination.
DataRaft / BarSuccessTechnical Lead & Sole DeveloperAcademic Analytics · University of Arizona

DataRaft / BarSuccess is an academic analytics platform for the James E. Rogers College of Law.

  • Replaced fragmented departmental spreadsheets with a centralized system for approximately 2,000 student records dating to 1990, including ingestion, ID-based merging, and validation checks.
  • Developed an LLM interface that translated requests into a restricted set of predefined functions for filtering, sorting, and updating large groups of records.
  • Developed commercialization strategy through administrator discovery and iterative demos, and presented the work at an AI and data conference.