Carla Prieto

Machine Learning Engineer

Hi there! You’ve reached my portfolio.

You can see my projects, experience and go ​to my contact page to leave a message after ​the beep.

Buildable Instagram UI Arrow Down Icon

Projects

In-house fine-tuned LLM ​to provide Airbnb ​recommendations.

Olivia : An AI traveler ​assistant

Implemented BERT and ​CNNs to categorize ​users phrases to ​applications in Alexa.

Mapping Alexa’s ​Utterances to intents

Implemented audio ​feature extraction ​techniques to apply ​multiple classifier ​models.

Music Genre Classifier

Buildable Instagram UI Arrow Down Icon

Experience

2022 - 2024 | Hundred X Machine Learning Engineer


2022 - 2024 | Hundred X

Machine Learning Engineer

  • Led project that analyzed customer comments to provide insights for both external clients and the internal insights ​team.
  • Co-designed and implemented dashboard for clients to visualize and understand customer comments trends.
  • Implemented a model to identify customer comments topics and trends over time using BERT.
  • Used GPT-3.5 API to name the topics customers were discussing.
  • Drove sentiment model improvements. Accuracy increased from 60% to 85% overall and 93% for English ​comments.
  • Designed a data pipeline to clean customer comments, apply the ML models, and populate the dashboard's data ​tables.
  • Collaborated closely with data science, insights, and customer teams to ensure project success.
  • Delegated data engineering tasks to the engineering team to streamline workflow and enhance efficiency.


Buildable Instagram UI Arrow Down Icon

2022 - 2024 | Hundred X Machine Learning Engineer


2019 - 2021 | Amazon

Software Engineer

  • Implemented feature based on machine learning model output to reduce time in manual data annotation by 5%.
  • Maintained machine learning models for real time prediction in a production environment.
  • Designed and developed data pipeline to deliver metrics to business intelligence.
  • Designed and implemented end to end tests and reduced deployment time by half.


2017 - 2019 | Spectrum Effect

Software Engineer

  • Helped implement data pipeline for feature extraction at large scale.
  • Collaborated with data scientists to implement machine learning models.
  • Implemented algorithm for localization using graph databases.


Buildable Instagram UI Arrow Down Icon

Skills

ML Frameworks

  • TensorFlow
  • Scikit-learn

Big Data

  • Apache Airflow
  • Kubernetes

Cloud Computing

Programming Languages

  • Python
  • Java


  • R
  • SQL


  • AWS
  • GCP

Data Analysis

Statistics

  • Pandas
  • NumPy
  • Tableau
  • D3
  • A/B Testing
  • Field Experiments
Buildable Instagram UI Arrow Down Icon

Education

UC Berkeley

2024

M.I. Data Science

Tecnologico de Monterrey (ITESM)

2016

B.S. Information and Communication ​Technologies

Buildable Instagram UI Arrow Down Icon

Contact

carla.prieto@berkeley.edu

Email Message Envelope
Phone

(206) 581- 4834

Socials

Buildable Instagram UI Arrow Down Icon
Buildable Instagram UI Arrow Down Icon

Olivia: An AI Travel Assistant

Olivia is a chatbot that gives AirBnB recommendations ​for your next vacation. Is backed by an in-house fine-​tuned LLMs. Comes in two flavors: Using Claude Sonnet ​or LLaMa 3.

Check it out!


oliviawebpagedemo.my.canva.site


Mapping Alexa Utterances to Intents

Voice-activated virtual assistants, like Amazon Alexa, have revolutionized human-computer ​interaction by responding in real-time to user instructions, known as “utterances.” These ​utterances are mapped to specific “intents,” which guide Alexa in executing the correct tasks.

In this project, I explored the use BERT (Bidirectional Encoder Representations from Trans- ​formers) along with another models to map utterances to intents.


Look at the full details in my repo:

github.com/Carla08/alexa_utt_intention_bert

Music Genre Classifier

A favorite of mine. In this project I learned how to extract and manipulate features from audio ​files, do some programatic audio chops, principal component analysis (PCA), and plug it all into ​common used ML models for classification.


I learnt about extracting and manipulating soundwaves, MEL Spectograms, harmonics and ​perceptruals among other audio features.


I also got to experiment with a bunch of models, from simple KNN and decision trees, to more ​robust ones like an SVM and a classifier using gradient booster.


Finally achieving an accuracy of ~0.90 for one of my first models.


Check the final presentation out! (contains pretty dope visuals!)