Catalin Caldararu

• Advised functional leaders on AI product strategy in an engagement focused on local AI.
• Originated, built, and shipped a private local AI system that powers data-driven acquisition decisions in a commercial niche, running at near-zero marginal cost per query.
• The system combines semantic embeddings, sentiment analysis, and real market data to identify, score, and rank acquisition candidates. The scoring informs the decision rather than making it.

Driving the portfolio's development until the final 13 inventions are handed off to the legal team for filing.

This was the R&D half of a single role, spanning research, product ownership, solution design, and delivery. This work is part of the sustaining innovation strategy for a commercial Hybrid AI platform featured at CES and Dell Technologies World. Its scope is to protect both the platform and its integrations with 8 applications across the organization, as they exist today and as they evolve.

• Initiated and led a Product Innovation workstream, originating, owning, and managing a 35-item Applied AI patent application portfolio as an R&D product: 22 applications filed, 13 in development.
• The inventions span 8 domains: Security, Privacy, Resilience, Efficiency, Accessibility, Personalization, Automation, and Agent-Native SDLC.
• Owned the portfolio's backlog and strategy as first-named inventor on all 22 filed applications, managing invention flow with Kanban.
• Incrementally secured $1M in funding by presenting business cases for all 35 concepts to Software Engineering and Legal leaders, becoming the top inventor at Dell's Romania site.
• As Technical AI PO for a research initiative focused on unreleased capabilities and future products, originated the requirements for all 35 AI product concepts, defining both their solution architectures and PRDs.
• Drove the portfolio's development, assembling invention teams of ~4 around each concept's domain from a global pool of 10 co-inventors, and refining the inventions with them and multiple attorneys.
• Personally built and shipped the core Hybrid AI services referenced in my patent applications.
• Coached colleagues on Applied AI and the invention process, and presented innovation topics to the team inheriting the workstream.

This was the development half of a single role, spanning research, back-end service design, development, testing, delivery, and monitoring. I was part of the core team of a Hybrid AI platform, and contributed as a developer to a separate Applied AI Desktop Application. Both are commercial products featured at CES and Dell Technologies World and ship to all Dell AI PC customers.

• Sole engineer for 6 OpenAI compatible APIs and inference services spanning two silicon vendors, including the default conversational services (SLM and ASR), which were, at release, the most performant Dell services for these targets.
• Contributed to inference services across NPUs, GPUs, and CPUs, spanning all 4 silicon targets (AMD, Qualcomm, NVIDIA, Intel).
• Built test suites measuring the accuracy, latency, and safety of candidate SLM and ASR models on task-specific benchmarks. These tests determined which models shipped.
• Delivered foundational components enabling internal teams and customers to build Hybrid and Agentic Edge AI applications across heterogeneous hardware and AI PC fleets.
• Facilitated Daily Scrums for a 5-person AI inference team, assigning tasks and clearing blockers.
• Resolved defects and dependencies while coordinating directly with hardware vendors and three other Scrum teams (ML, DevOps, and a downstream service consumer).
• Negotiated scope with senior leadership, including Engineering and Product Leaders from Dell's CTO organization.
• Reported progress to internal and external stakeholders, including directors and Fellows.
• Monitored service quality across successive hardware, driver, and dependency versions, spotting degradation trends early and re-optimizing before they were released.

Worked in a 5-person feature team in a scaled modernization program spanning 10+ teams. This has involved back-end service design, development, testing, delivery, and monitoring.

• I was an early adopter of AI coding tools, which eventually led me into more AI-driven work.
• Designed and developed 3 full-stack modules comprising of 6 .NET APIs and 2 React front-ends, with 3 APIs also being consumed by other Scrum teams.
• Facilitated Daily Scrums, cleared blockers, and served as the sole integration engineer for the team's releases.
• Collaborated closely with the Engineering Manager, Program Manager, other leaders, and many other internal stakeholders.
• Ensured deliverables met non-functional requirements across automation, security, accessibility, observability, secret management, and performance.
• Onboarded 4 developers with various levels of experience, enabling the team to take on more work.
• Prevented technical debt through static code analysis, refactoring, code reviews, and pair programming.
• Interacted with customers (internal stakeholders) and supported cross-team UAT/E2E defect triage.
• Monitored service quality and delivered targeted improvements.