Error Analysis and Fine-Tuning of Google Gemini.

Problem Setter in Python. Designed coding questions in Python (with solution and test cases).

As a Freelancer (Part-time consultant), I work closely with the course mentors and assist the students in clarifying their subject matter queries in Python, C++ and SQL.

Designed a Dashboard to track the performance of Job-to-CV Matching algorithm running on Naukri platform (along with applies and recruiter actions) using Plotly Dash.Manual inspection of the data and cleaning it using different methods based on intuition (trend).Performed exploratory data analysis on the job posting data of Naukri Gulf.Made a report on the current Job-to-CV Matching algorithm implemented on Naukri and Naukri Gulf and identified the drawbacks of the current model implemented in Naukri Gulf.Applied modified version of Naukri ARS (Apply Relevance Score) model on Naukri Gulf.

Mentored a group of 5 students on GANs and CycleGANs.Prepared mock Data Science Interview questions based on different topics.Mentored a group of 6 students for a blog of Naive Bayes to be posted on behalf of KDAG.Contributed to ML workshop on the topic of Principal Component Analysis (Mathematics as well as code explanation with Quiz).Mentored Junior members of KDAG in CSR team for Kharagpur Data Science Hackathon.

Contributed to the ML workshop conducted by KDAG in preparing lectures slides and quizzes on ensemble models. Hosted the Summer AI Challenge event sponsored by Elastic on the Dare2Compete portal and successfully achieved 800+ registrations.

Read research papers on GAN, CycleGAN and LPTN. Used CycleGAN and LPTN for converting RGB images of vKITTI to KITTI along with visualizing the performance of the model using Tensorboard.

Prepared multi label classification models for image quality detection of images of cars and used CycleGAN to generate more images of some paticular classes.Used DeepMARC and Detectron2 models for Carparts segmentation.Performed Panorama stitching to generate 3D images of cars.

Extracted Twitter data using Twitter API by filtering retweets.Analyzed Twitter data using Topic models.Performed emotion mining on the textual data and data visualization to represent the relationships between the topics and emotions