Summary
Work History
Education
Skills
Projects
Additional Courses
Accomplishments
Volunteered Services
Languages
Hobbies and Interests
Timeline
Generic
Sam Kingsley

Sam Kingsley

Bangalore

Summary

Goal oriented Machine Learning engineer, with a degree in Aerospace Engineering, Meticulous individual brings expertise in testing, validating and improving models in terms of accuracy and computational time for complex use cases. Builds models for different use cases with wide variety of available data.

Work History

Machine Learning Engineer

Pinaca Technologies Pvt. Ltd.
Chennai
09.2023 - Current
  • Provided backend support with trained models and alterations for real-time predictions to build a web-application for the same.
  • Evaluated model performance by creating metrics such as accuracy, precision, recall, F1 score and AUC.
  • Optimized existing models to improve accuracy and reduce computational time.
  • Conducted research on new algorithms in the field of deep learning.
  • Performed hyperparameter tuning for improving model performance.
  • Worked closely with clients to define requirements for ML products.
  • Reviewed code changes made by team for accuracy and relevance.
  • Communicated complex machine learning solutions and concepts to diverse group of business clients in clear, concise language.

Education

Post Graduate Diploma - Data Analytics and Machine Learning

Imarticus Learning Institute
06.2023

Bachelor of Technology - Aerospace Engineering

Karunya Institute of Technology And Sciences
05.2022

Skills

  • Coding: C, Python, C
  • Machine Learning and Data Analytics
  • Deep Learning
  • Model Development
  • Database management using SQL, MongoDB
  • Vector Databases - Qdrant
  • Data Visualization using Power BI and Tableau
  • Flow visualization using ANSYS
  • MS Office
  • Adaptability and Flexibility
  • Well organized
  • Time Management
  • Problem Solving

Projects

  • Face Recognition for videos in real time - Two shot model

      - Processing input video to extract faces and embed them using a model to check cosine score with source faces. The process metrics were improved along with computational time using multiprocessing. Databases such as MongoDB and Qdrant vector database were used in this project. 

  • Document search by contextual words - Contextual Semantic Search

     - Document search included cosine score lookup from Qdrant vector database and then results were re-scored using a personalized key-word search algorithm based on json structures to store the data of keywords in the local folder temporarily.

  • Financial fraud detection using GNNs

     - Built a GNN model based on the Graph that was built based on unstructured classification labeled data. The model was semi-supervised and inductive based on a heterogenous and dynamic graph. The model architecture and data pre-processing was designed to predict financial frauds with high accuracy.

  • Prediction of usability of a turbofan engine using Machine Learning Algorithms

     - The dataset used for this project is from NASA's official dataset collection. The task was to predict the lifecycles of each components in aircraft engines to reduce maintenance costs. The dataset included data form 25+ sensors which were recorded as individual features. Data was pre-processed and Principal Component Analysis was used to reduce the dimensionality. Then the classifier was tested with different models such as Logistic Regression, Support Vector Classifier, KNN Classifier, Decision Tree Classifier and Random Forest Classifier. 

  • Estimation of fuel using MATLAB - Chandrayaan-III, ISRO, ISTRAC, Bangalore

     - Given a lunar trajectory and itinerary of a lunar lander separated from an orbiter in the low lunar orbit, estimation of fuel was done for the lunar lander to land on the surface of the moon. If the fuel in the reservoir was more than necessary, the chances of crash landing due to excess mass would increase. On the contrary, if the fuel in the reservoir was less than necessary, the chances of crash landing due to lack of deceleration force would increase.

Additional Courses

  • Python Programming - Coursera
  • Database Management (Python and SQL)
  • Data Manipulation from Internet (Python)
  • Search Engine Development
  • Twitter Developer Access

Accomplishments

  • Building a drone, programmed for remote controlling using Arduino
  • Designing a CubeSAT using Fusion360 and SolidWorks software

Volunteered Services

  • National Service Scheme (NSS) (Karunya Institute of Technology and Sciences) - Organized a Health Camp in a nearby village and Co-ordinated in an Environment awareness program.
  • Lead a blood donation campaign in university
  • Directed an event for the music club in university

Languages

  • Tamil - Mother Tongue
  • English
  • Hindi
  • German

Hobbies and Interests

  • Programming
  • Astronomy
  • Reading
  • Structural Advancement of Materials
  • Travelling
  • Playing Music - Guitar, Piano
  • Sports

Timeline

Machine Learning Engineer

Pinaca Technologies Pvt. Ltd.
09.2023 - Current

Post Graduate Diploma - Data Analytics and Machine Learning

Imarticus Learning Institute

Bachelor of Technology - Aerospace Engineering

Karunya Institute of Technology And Sciences
Sam Kingsley