Erick M.

Key Skills

Software

Google Cloud Vertex AI

Top Subject Matter

Animal Species Identification

Road Sign Recognition - Autonomous Driving

Food Image Classification

Top Data Types

Image

Text

Top Task Types

Classification

Text Generation

Freelancer Overview

Animal Image Recognition - Custom Dataset Labeling & Training. Brings 3+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include TensorFlow and PyTorch. Education includes Bachelor of Science, University of Embu (2023). AI-training focus includes data types such as Image and Text and labeling workflows including Classification and Text Generation.

Labeling Experience

Bigram Language Model - Swahili Text Data Curation & Labeling

TextText Generation

I trained a character-level language model on Swahili text data to enable next-character prediction and text generation. The role required curating and preparing underrepresented language data, then labeling and organizing it for efficient supervised training. The final model contributed towards supporting Swahili NLP research and applications. • Collected and cleaned Swahili text corpora from various sources • Labeled and segmented datasets for training and validation • Addressed noise and inconsistencies in the data • Used PyTorch for model implementation and evaluation.

Not specified

Food Vision Mini - Food Image Dataset Labeling & Fine-tuning

ImageClassification

I implemented a transfer learning model using EfficientNet to classify 101 food categories based on labeled food images. My responsibilities included curating and labeling a diverse food image dataset and validating data quality to maximize classification accuracy. This supervised learning project helped improve food recognition capabilities in AI-based applications. • Labeled and organized images into 101 distinct food categories • Participated in dataset cleaning and outlier removal • Conducted data checks for label consistency • Leveraged labeled dataset to fine-tune the EfficientNet model.

Not specified

Road Signs Classification - Image Labeling & Model Training

ImageClassification

I developed a classification model for road signs utilizing a labeled image dataset with 44 distinct road sign types. The process included annotating and separating images for proper category representation to train and validate the model. This ensured the resulting deep learning model could be applied effectively to autonomous driving safety systems. • Sorted and annotated images by road sign category • Conducted data validation and relabeling for accuracy • Managed dataset balance for all 44 classes • Evaluated labeled data impact via model performance reviews.

Not specified

Animal Image Recognition - Custom Dataset Labeling & Training

ImageClassification

I created a CNN model to classify 90 animal species from images, developing a custom dataset for model training. The data labeling process involved categorizing images into their respective animal species to prepare data for supervised learning. My work contributed to achieving an 88% classification accuracy for the animal image recognition task. • Collected and organized images into proper classes for labeling • Labeled and validated category correctness per image sample • Addressed class imbalance and ensured high-quality data input • Used model training feedback to refine and re-label as needed.

Not specified

Education

Work History