Led linguistic operations for Siri’s Spanish-language development, coordinating ASR (Automatic Speech Recognition), TTS (Text-To-Speech), and ITN (Inverse Text Normalization) improvements through agile methodologies.

​

Main responsibilities included:
 

• ASR: ticket triage and debugging of speech recognition outputs.

• TTS: annotation, training, and debugging for text-to-speech quality and accuracy.

• ITN: reviewing and updating normalization rules to ensure consistent text and number formatting across modules.
   

Managed training and onboarding for new linguists, identified workflow optimizations, and implemented process improvements that increased accuracy and efficiency across teams.

​

Flagged critical data curation issues for cross-functional groups, enhancing product reliability and visibility of linguistic quality within the organization.
 

(Based on a LinkedIn recommendation by former supervisor Abe S., PdM Ops @ TPS, Aug 2022)

I worked on semantic annotation for Amazon’s search engine, contributing to the improvement of query understanding and relevance modeling.

Volunteered with Omdena as Project Specialist, participating in and mentoring a project based in India that developed an AI tool for linguistic and psychometric assessment of soft skills throughout both face-to-face and online meetings.

Designed and developed a visual-cognitive system for teaching and learning language through semantic mapping.
 

Inspired by the same principles that underlie LMs (Language Models) —context, association, and pattern recognition—, it helps learners infer relationships between words instead of memorizing them.

​

[→ Learn more here ]

Recruited by the U.S. Department of State, in collaboration with the U.S. Embassy in Spain, to design, coordinate, and teach a custom syllabus for gifted teenagers with limited resources at FIS (Fundación International Studies).

​

The program fostered entrepreneurship, public speaking, democratic values, human rights awareness, and technological literacy in English —including introductory modules on AI and programming fundamentals.

Implemented the Change Makers and Technovation Challenge curricula within the syllabus, enabling all students to identify problems in their environment and design practical solutions, develop their own apps and business projects as a team —discovering which role best fit them— and coordinating through Scrum meetings and agile milestones to help them become familiar with iterative project development.

Supervised and coordinated Fulbright Teaching Assistants and volunteer teams among other management duties.

Outcomes:

​

• One of the all-girl teams I coached at Go! American English reached the Quarter Finals at the Technovation Challenge in Madrid (2019).

​

• I was awarded a Certificate of Appreciation by the US Embassy in Spain "in recognition and deep appreciation of the significant amount of energy, resources, and dedication that you provide to the US Embassy-FIS English Access students."

​

• Over the years, I have stayed in touch with several of my former students and their mothers from both programs. Some have even come to visit me —sometimes staying for a few days— and I've supported others through advice, assistance with applications procedure, or recommendation letters for scholarships and new opportunities.

Collaboration in linguistics research projects as Honorific Research Fellow at the Linguistics Department of UCM related to:

• Epistemic modality

• Automatic generation of dialogues

Worked on the development and maintenance of Google Assistant for Spanish (es-ES).

​

Main tasks included:

​​

• NLG (Natural Language Generation) and NLU (Natural Language Understanding) annotation, including semantic, syntactic and morphological tagging.
   

• Deriving and curating NLP (Natural Language Processing) data for new and ongoing projects.
   

• Support during Google Assistant product launch events, such as the release of the Pixel phone.

Worked as part of the R&D team developing Bixby, Samsung’s voice assistant, for the Spanish (es-ES) market.

I was responsible for the G2P (Grapheme-To-Phoneme) module within ASR ensuring accurate phonetic transcriptions and pronunciation modeling.

My main tasks included:
 

• Designing and maintaining grammars to create representative test cases for ASR performance.
   

• Triaging and analyzing ASR error results to identify root causes and assign appropriate fixes.
   

• Supporting the AE (Audio Engineering) team with quality control and phonetic validation tasks.

Initiated and led a collaborative research project between the Computer Science Department at UPSAM (Universidad Pontificia de Salamanca en Madrid) and the English Linguistics Department at UCM (Universidad Complutense de Madrid), based on the EAGLES standards for lexical tagging and annotation.

​

The project later evolved into a joint initiative with UB (University of Bremen) thanks to Dr. Zamorano-Mansilla —who directed the project— focusing on the development of an ontology for an automatic language generator developed by UB (KPML).

Eventually, the team focused on the behavior of certain adjectives and their "intensional" modification within nominal groups —resulting in two conference papers and one publication.

Conference papers:

• AJL's International Congress. XXV Encuentro: Modificación intensional en el grupo nominal en inglés y en español (2010). 

• I International Congress: languages and their literatures (AJIF-UCM): La modificación intensional en el grupo nominal en inglés y español: un estudio contrastivo aplicado a la computación. (2007)

​

Publication:
 

• Gimeno Miró, Esther. Modificación intensional en el grupo nominal en inglés y en español, in Interlingüística XXI, Valladolid, 2011.

​

This earlier research project laid the theoretical foundation for my later work in applied AI, bridging linguistic theory, lexical semantics, and computational modeling.

[→ Learn more here ]

Worked on the MULTINOT project as a PhD candidate at UCM, researching and designing a coding scheme for annotating an English–Spanish corpus in the field of verb tense analysis.

Designed and developed a visual-cognitive system for teaching and learning language through semantic mapping.
 

Inspired by the same principles that underlie LMs (Language Models) —context, association, and pattern recognition—, it helps learners infer relationships between words instead of memorizing them.

​

[→ Learn more here ]