Catalin Caldararu

I spend one‑third of my time leading a 35‑initiative Applied AI Intellectual Property (IP) portfolio focused on Agentic Edge AI and Hybrid Edge‑Cloud AI, owning its strategy, prioritization, and lifecycle as a product. To date, this work has produced 12 USPTO patent filings (10 as first inventor, 2 as co‑inventor) and 8 internally recognized inventions. Key responsibilities include:

• Initiating new ideas, contributing directly, and coordinating with a core team of 4 repeat inventors and multiple legal specialists.
• Driving the patent development process from opportunity discovery through concept refinement, business‑case development, technical design, and iterative improvement.
• Defining portfolio, patent, and product strategies, aligning work with the company's technology strategy, and discontinuing initiatives when necessary.
• Originating AI product concepts that can create competitive advantage, support legal defense, or generate revenue.
• Applying agile product‑ownership principles to prioritize the IP backlog and estimate key product metrics, including current value, unrealized value, time to market, and innovation potential.
• Creating detailed technical architectures that enable independent execution by engineering teams.

This work has refined my product development and design‑thinking skills through hands‑on application in a Fortune 500 engineering environment, working primarily with engineering and legal teams rather than traditional end‑user‑facing contexts.

I spend two‑thirds of my time as a core contributor to multiple on-device Small Language Model (SLM) and Automatic Speech Recognition (ASR) services used in Dell's AI PC platform, which was showcased as a flagship AI initiative at major industry events (CES, DTW, GTC, etc.). This work serves as a foundational capability for multiple Agentic Edge AI and Hybrid Edge‑Cloud AI initiatives. Key responsibilities include:

• Serving as the primary contributor to six SLM‑based conversational AI services and three automatic speech recognition (ASR) services. Two services were 0‑to‑1 developments that established the foundation for multiple future initiatives. Several became either the only Dell solutions in their hardware class or the most performant at the time of delivery.
• Acting as lead developer and primary contributor for all on-device Conversational AI services delivered on Dell devices through a partnership with a leading AI hardware vendor, collaborating directly with external stakeholders and Dell's CTO organization.
• Designing and implementing on‑device AI inference pipelines for open‑source ONNX models (e.g. Hugging Face-hosted models) using ONNX Runtime, enabling efficient execution across heterogeneous CPU, GPU, and NPU hardware.
• Evaluating AI model performance, behavior in specialized tasks (e.g., MMLU), and robustness against adversarial inputs.

This work has grounded my product thinking, enabling me to execute, evaluate and incrementally improve software products rather than just conceptualize them.

Built a strong foundation in .NET Web API development, GitLab CI/CD automation, testing, and application observability, then progressed to delivering multiple microservices end‑to‑end, mentoring three interns, onboarding four engineers into product teams, and collaborating with engineers from eight other teams and multiple business stakeholders.

Contributed to 15 educational projects helping students develop technical and soft skills and transition into IT roles. Also supported internal and partner recruiting through candidate triage and early‑stage HR interviews.