Training to Implementation: My Data Science Internship Journey at Pickl.AI

Introduction:

My data science internship at Pickl.AI has been a transformative experience, spanning two distinct phases: a month of intensive training and learning, followed by the exhilarating phase of project implementation. Little did I know that the second phase would bring me face-to-face with an exciting project aimed at building a model to detect payment fraud. This is the story of how I navigated through these phases, embracing challenges, and emerging with newfound skills and confidence.

Phase 1: A Month of Learning and Upskilling The first month of my internship was dedicated to training and upskilling. I was introduced to various data science concepts, programming languages, and tools such as Tableau and SQL. The training sessions covered everything from data preprocessing and feature engineering to advanced machine learning algorithms. Although the learning curve was steep, I felt exhilarated to absorb the vast knowledge imparted by seasoned data scientists.

Real-World Case Studies and Guest Lectures: During this training phase, real-world case studies provided valuable insights into applying data science techniques to solve complex problems. These case studies showcased the relevance of data science in different industries and piqued my interest in real-world applications.

Guest lectures by industry experts added a layer of practicality to the theoretical concepts I had learned. Their experiences and expertise inspired me and instilled in me the desire to apply my knowledge to solve real-world challenges.

Phase 2: Unveiling the Project and Project Implementation As the training phase concluded, I was filled with anticipation to apply my skills to real projects. To my delight, I was assigned to work on a challenging project: building a model to detect payment fraud. The moment I learned about my project, I was thrilled and eager to take on this new challenge.

Project Kick-Off: The project kick-off marked the beginning of the implementation phase. I was facing problem in dealing with outliers but I calculated various performance matrics.Because of Reverse engineering I get to know what to do with outliers. Every week I had a project discussion session in which I could share my doubts and know how to approach the project.

Navigating Challenges and Gaining Confidence: As I dived into the project, I encountered various challenges. Data preprocessing and handling imbalanced data were among the initial hurdles I had to overcome. However, I tackled each challenge methodically and developed a better understanding of the project's intricacies.

Model Building and Iteration: Building the payment fraud detection model required thoughtful feature engineering and selecting the appropriate machine learning algorithms. I used Logistic Regression, Decision Tree and Random Forest as my ML algorithms.

However I struggled with Hyperparameter tunning part as it requires high GPU so I completed my project till calculating performances matric. And I get to know that recall would be the best performance matrics to evaluate the performance of my model.

Final Deliverables and Sense of Accomplishment: After weeks of hard work and dedication, the moment of truth arrived as I finalized the model and presented our findings. The sense of accomplishment was overwhelming as I showcased the successful detection of payment fraud through data science with an accuracy above 99%.

Conclusion: My data science internship journey at Pickl.AI has been a remarkable experience of learning, growth, and personal development. The transition from training to project implementation provided the perfect platform for me to apply my knowledge and skills to a real-world problem.

I am grateful for the mentorship, support, and encouragement I received from the Pickl.AI team throughout my journey. The challenges I encountered, coupled with the triumphs, have instilled in me a sense of confidence in my abilities as a data scientist.

As I bid farewell to my internship, I carry with me a newfound passion for data science and the determination to continue exploring its boundless possibilities. The experience at Pickl.AI has ignited a spark within me, motivating me to pursue a fulfilling career in data science, armed with the tools and knowledge I gained during my internship.