Summary
Work History
Education
Skills
Websites
Accomplishments
Disclaimer
Languages
Timeline
Generic

Satyam Harikrishan Kshirsagar

Pune

Summary

Data scientist with 1.5+ year of experience in statistical analysis, machine learning, deep learning techniques, and data visualization tools to solve business challenges. Hands-on experience in forecasting methodologies to anticipate future trends along with computer vision techniques. Passionate about continuous learning and staying updated with the latest advancements in the field of data science and AI.

Work History

Data Scientist

Rubiscape PVT .LTD
  • Objective: To detect original and duplicates bottles of Hair oil using CV and ML
  • Use YOLO for the bottle's upper cap and bottom
  • Identify key parameters from the bottom part of the bottle
  • Determine key positions for these parameters and integrate them with the metadata for each bottle
  • Use Optical Character Recognition to detect key terms from the bottle's side information
  • Build a classification and object detection model for this data
  • Develop interactive dashboard on streamlit and android app to collect the data using flutter and django
  • Client has found almost 50 Cr of counterfeit bottles are there in the market and using this, there is almost 5 % of revenue growth and big business impact
  • Technologies: Python, CNN, YOLO, OCR, Flutter, Django

Data Scientist

Rubiscape PVT .LTD
  • Objective: To verify whether the invoice is valid or not by Detecting key elements from invoice using Computer vision
  • Developed Prediction pipeline which runs every night includes scrapping PDFS of invoices from portal, converting PDFs to images, detecting key parameters from invoice like stamps, invoice no, IRN no, Important symbol etc
  • Also runs OCR model to get different key information from invoices
  • Valid invoices data get stored on snowflake and links for PDFS are stored at Amazon S3 buckets
  • Experimented with augmentation techniques, Different models, and models parameters so as to capture different shapes and colors of stamps, and models performance improved from 72 % to 87 %
  • Technologies: Python, CNN, YOLO, Faster RCNN, OCR, Git-hub, Amazon S3, snowflake, Web scrapping

Data Scientist

Rubiscape PVT .LTD
  • Objective: To Detect Anomaly on Distributer and retailer level using ML techniques
  • It has been found that to get benefits of offers, and to complete the given target, Few distributers shows anonymous behavior in sales which affects supply chain, planning and procurement highly
  • Developed an ensemble model giving Dates of possible anomalies for distinct distributers and using models output, statistical method like hypothesis testing, EDA Found out most probable Distributer retailers pairs to make anomalies
  • Build Interactive Dashboard using Streamlit
  • Technologies: Unsupervised learning, Snowflake, EDA, Streamlit

Data Scientist

Rubiscape PVT .LTD
  • Objective: To forecast the sales quantity for the next period so that the organization can have inventory for the upcoming demands in advance
  • The company is losing customers due to the unavailability of products (facing a stock-out situation)
  • The forecasted sales quantity will help to maintain sufficient inventory which will be enough to meet future customers' demands
  • Sudden demand from customers is fulfilled by importing goods via airway which incurs huge costs
  • The forecasted values will give time for goods that can be imported via sea which drops the cost to 20%
  • Forecasted sales quantity is found to be resembling real-time data for 75-80% of Items
  • Technologies: Python, PowerBi, Deep Learning, Machine learning, Arima, Sarima

Data Scientist

  • Using the NLP Hugging Face transformers library (RoBERTa based) sentiment scores of Economics News are gathered and appended in Clients' data as sentiment-score
  • Then for the whole data set LSTM model is applied
  • So that future prices can be predicted by considering the sentiments of the market and risk for investor
  • By using an exogenous factor of Sentiment, error in prediction is reduced and on a broader scale, we could drastically reduce traders' risk
  • The model with a sentiment score is making 3% fewer errors as compared to that without sentiment
  • Technologies: Python, Transfer learning(Roberta-Based Transformer), NLP, Deep Learning

Education

Post-graduation - Data Analytics and Machine Learning

Imarticus Learning
Pune, Maharashtra
10.2024

Bachelor OF Engineering -

Sinhgad College Of Engineering, Pune University
Pune, Maharashtra
01.2021

HSC -

Apte College
Pune, Maharashtra
01.2017

Skills

  • Languages and Databases: Python, Snowflake, SQL
  • Data Visualization: Power BI, EDA , Streamlit
  • Statistics and Probability: Hypothesis Testing, Descriptive Statistics, Making Inferences from Data
  • Machine Learning: Supervised and Unsupervised Learning like SVM, Decision Tree, Logistic regression, PCA, Hierarchical and K-means clustering, KNN , Random Forest, XGBOOST, GBMs
  • Deep Learning: Feed forward network, RNN , LSTM, GRUS, Tensorflow, Keras
  • Time Series Forecasting: Statistical Methods, ML, DL, and Transformer-Based Models
  • NLP: TF-IDF, Word Embeddings,Encoders, Decoders, Transformer-Based Architectures
  • Computer Vision: CNN, Object Detection and Classification, YOLO, Faster R-CNN, Data Wrangling
  • Optimization techniques: Adam, L1 L2 regularization, adadelta, Binary and Categorical Cross entropy
  • Basics of Amazon AWS (lambda function, S3)
  • Generative-AI

Websites

Accomplishments

    Won rising star award for displaying outstanding commitment to the role., 04/24

Disclaimer

I hereby confirm and verify that all the information mentioned here and I take full responsibility for its accuracy and authenticity.

Languages

  • English
  • Marathi
  • Hindi
  • German

Timeline

Data Scientist

Rubiscape PVT .LTD

Data Scientist

Rubiscape PVT .LTD

Data Scientist

Rubiscape PVT .LTD

Data Scientist

Rubiscape PVT .LTD

Data Scientist

Post-graduation - Data Analytics and Machine Learning

Imarticus Learning

Bachelor OF Engineering -

Sinhgad College Of Engineering, Pune University

HSC -

Apte College

Similar Profiles

  • Randhir Kumar Prasad

    Randhir Kumar PrasadRandhir Kumar Prasad

    Stacker at Engineering and Allied Services Pvt. Ltd, Lift & Shift India Pvt. Ltd, Roadwings International Pvt. Ltd, Western Carriers India Pvt. Ltd, Tractor India Pvt. Ltd, Transworld Terminal Pvt. LtdStacker at Engineering and Allied Services Pvt. Ltd, Lift & Shift India Pvt. Ltd, Roadwings International Pvt. Ltd, Western Carriers India Pvt. Ltd, Tractor India Pvt. Ltd, Transworld Terminal Pvt. Ltd

  • SARVESH SHRIYAN

    SARVESH SHRIYANSARVESH SHRIYAN

    Application Consultant /Senior Analyst at Rtv Tech Pvt Ltd (Subsidiary of Rupeeseed Pvt Ltd)Application Consultant /Senior Analyst at Rtv Tech Pvt Ltd (Subsidiary of Rupeeseed Pvt Ltd)

  • VIJAYA RAGUNATH V

    VIJAYA RAGUNATH VVIJAYA RAGUNATH V

    MANUFACTURING PROCESS ENGINEER at MINI-CIRCUITS INDIA URJITA PVT LTD (Formerly URJITA ELECTRONICS PVT LTD)MANUFACTURING PROCESS ENGINEER at MINI-CIRCUITS INDIA URJITA PVT LTD (Formerly URJITA ELECTRONICS PVT LTD)

  • Pratik Chafle

    Pratik ChaflePratik Chafle

    Data Engineer Manager at SaksoftData Engineer Manager at Saksoft

  • Vaibhav Suryawanshi

    Vaibhav SuryawanshiVaibhav Suryawanshi

    UI Developer at Xcaliber Infotech Private LimitedUI Developer at Xcaliber Infotech Private Limited

CREATE PROFILE
Satyam Harikrishan Kshirsagar