Jacinta Clovis

Driven and versatile ML Engineer with a proven track record in developing and deploying machine learning solutions that deliver meaningful insights and exceptional end-user experiences. Specialist in Python libraries NumPy, Pandas, Matplotlib, Scikit-learn, TensorFlow, PyTorch, SciPy, Keras, and others. Fluent in Javascript, React, Redux, and CSS; accomplished in AWS Lambda, API Gateway, S3, and CloudFormation. Success in varied self-lead development projects including human activity recognition modeling, inertial measurement data analysis, front-end dashboard development, as well as a keen understanding of deep learning algorithms. ML projects include a multi-class classification traveling model and the piloting of a tripping and talking model in Python. Clear track record of flourishing in new competencies when required for new projects, and impressive history of succeeding with every new responsibility. Experience in algorithm development, with employment of segmentation and feature selection methods. Equipped to design, integrate, and optimize models that effectively address complex business challenges. Avid student of new languages in my spare time, always with the ambition of successful application in new projects.

Whilst at Mafic, I developed numerous ML models - time series human activity recognition models for construction workers. The most complex project that I lead was a multi-classification traveling model, used to identify when a person is walking up or down the stairs, using a multi-floor hoist, or simply just walking. The model required me to develop a variable window length segmentation algorithm to classify the activities. Other models for which I piloted the development were to identify talking, tripping, and when the IMU had been dropped. This included model integration into the pipeline using MLFlow, providing me with hands-on experience with CI/CD and DevOps.
 

I took over the development of the web app in a front-end role where I learned JavaScript and React. I also integrated APIs using AWS Lambda, API Gateway, and S3. In this project, I developed dashboards for users to see their ML data predictions for their activities and a database web app used to manage tradesmen for supervisors and managers. I wrote APIs that requested data from Dynamo DB and S3 in AWS, and managed those APIs in API GateWay, CloudWatch and CloudFormation. Other projects within this area also included using MQTT for live data dashboarding for other sensors.

I utilised a feature generation library and developed a selection method that is now used across all ML projects the company undertakes. This method would generate thousands of features which would then be whittled down to less than thirty, all while remaining efficient in minimising memory usage. This largely automated the process of developing ML models within the company.
 

I transitioned from a summer intern to a part-time employee whilst studying for my Master's, ultimately becoming a full-time employee who plays an essential role in the company. I've worked on many projects and diversified my skills, not only refining the analytical skills I had in Python but also learning JavaScript and AWS services.

MSc Physics from the University of Southampton