Felipe Chavarro Polanía

Profile

Enterprise technology professional with HPE and Hewlett-Packard experience since 2012 across AI service offerings, product strategy, customer delivery and hybrid cloud architecture. My work also includes AI-powered process automation and software-factory development. Independent technical work includes Python research tools, LLM-judge auditing and compute-governed experimentation. I connect customer needs and product decisions with explicit evaluation criteria and reproducible evidence.

Professional experience

Hewlett Packard Enterprise / Hewlett-Packard

August 2012 to present

WW Service Offer Manager AI / AI Expert

January 2025 to present

• Work across AI product development, service offerings and customer enablement, connecting AI capabilities with enterprise needs.
• Define product strategy and communicate roadmaps for AI initiatives, including LLMs, AIOps and intelligent automation.

Earlier roles within HPE

WW Business Segment Manager; Practice Delivery Leader for LATAM; Cloud Engineer / Hybrid Cloud Solutions Architect; Lead Technical Consultant.

• Managed service lifecycle planning and cross-functional coordination across global teams.
• Shaped delivery strategies with presales teams, aligning customer expectations, resources and regional opportunities.
• Conducted cloud maturity assessments, solution architecture, cost-benefit analysis and technical demonstrations.
• Worked with executive stakeholders on technology strategy, migration roadmaps and enterprise delivery.

IBM · IT Specialist / Presales Engineer

2011 to 2012

Technical work across AIX, Linux, Citrix, VMware, virtualization and enterprise infrastructure.

Udacity · Tech Mentor

2021 to 2025

Reviewed code and projects and mentored learners in cloud computing, data science and programming.

AI-powered process automation

Software factory

Created an AI-powered software factory that automates software development and enables Product Management to build robust prototypes in weeks, accelerating time-to-market and time-to-value.

Independent voice AI work

Audiobook production and voice testing with ElevenLabs

Since 2024

• Produced more than 100 audiobooks in English and Spanish.
• Tested ElevenLabs v2 voices, checking generated audio for artifacts.

Selected open-source work

Governed AI Systems

Python research toolkit covering fractional-factorial recipe discovery, compute-governed experiment promotion, and identity-aware LLM-judge auditing. Implements inspectable, deterministic decisions from declared policies and structured evidence.

github.com/felipechavarropolania/governed-ai-systems

Repository examples are synthetic demonstrations; empirical evidence is provided separately in the preprints below.

Public research

• Staged Factorial Screening for Budget-Constrained Micro-Pretraining. Sole-authored arXiv preprint, 2026. A bounded methods study across 613 experiments, repeated checks, and continuations on two hosts. arXiv:2606.05186
• Small Experiments, Cheaper Decisions: A Case Study in Staged Promotion for Micro-Pretraining. Sole-authored arXiv preprint, 2026. Auditable promotion rules and staged compute allocation, including dual-host 12-hour confirmation. The full protocol reports 169.2 training GPU-hours. arXiv:2606.11387

Preprints are listed as preprints, without implying peer-reviewed publication. See each source for methodological limits.

Education and languages

Master of Science · University of Liverpool · 2013

Telecommunications Engineering · Universidad Piloto de Colombia · 2008

Spanish: native · English: full professional proficiency

Areas of focus

AI product development · AI-powered process automation · Agent harness engineering · AI evaluation · Experimental design · Solutions architecture · Hybrid cloud · Customer enablement · Cross-functional coordination