Siddharth Chaudhary Email and Phone Number
• 5+ years Industrial experience in driving business value using advanced Data Science/Analytics, Machine Leaning, Artificial intelligence techniques by leveraging on python, pyspark, SQL and applying deep learning methods to solve business problems using TensorFlow, Keras, Sklearn libraries.• Experience of implementing various NLP/ML models: RNNs, ANN, LSTMs, Sequence models, text-numeric-categorical multi input functional hybrid models, CNNs for text classification and XGboost, SVM, KNNs, Logistic Regression.• Effective collaboration with stakeholders, understanding their business needs to provide solutions.• Guiding new joiners and mentoring interns to manage their daily tasks.• Developed a Very Deep CNN model to predict whether medical charts contain evidence of the Disease, which potentially helped close over millions dollars gap in savings by identifying missing diagnosis.• Created a text-numeric-categorical complex NLP/ML functional tensorflow model with different modalities multi-inputs layers, accommodate structured & unstructured input data to classify provider enablement.• Developed and deployed a text extractor framework to fetch text sections from multiple documents to generate contract for providers using docx library, eventually replacing the legacy systems• Built text classifier using LSTM to identify the over/under payments for the claims, managed to save huge amount.• Implemented XGboost risk model for providers assessment with incorporation of mlflow for hyperparameter tuning.• Created few end to end scalable ETL pipelines using pyspark and python which eventually reduced the overall risk scores prediction and delivery time by almost 50%.• Experience of writing complex SQL queries against huge databases.• Built Tableau dashboards working with BI team. Built webapps using flask, python, html bootstrap. Also designed and developed more than 20 web dashboards using python, plotly dash, dash components including covid dashboard.• Worked with advanced tools: Jupyterlab, Databricks, IQstudio, Spyder, RStudio, Tableau, DBeaver, Airflow, GIT.• Completed Certification courses in Python, Pyspark, SQL, LSTMs, Deep Learning, NLP specialization, R, Dashboards.• Carried out a research on forecasting of solar radiation using Machine Learning (ARIMA, TBATS models), further predicted solar electricity generation using the forecasted solar radiation (Thesis).
Unitedhealth Group
View- Website:
- unitedhealthgroup.com
- Employees:
- 100135
-
Principal Data ScientistUnitedhealth GroupDublin, Ie -
Senior Data ScientistUnitedhealth Group Feb 2024 - PresentDublin, County Dublin, Ireland -
Data Scientist 2Unitedhealth Group Dec 2021 - Feb 2024Dublin, County Dublin, Ireland• Implemented a Very Deep CNN model (Inspired by research paper published by Facebook) to find evidence of a condition in medical charts. This architecture tokenizes chart text sequences then generates the Word2Vec word embeddings and passing it to a tf.keras.layers.Embedding layer then fits a deep 1D CNN model. Uses max pooling after every convolutional layer with a small amount of dropout. Also applies the dropout between the fully connected layer and final output layer. Uses paired convolutional layers with the 64 filter size.Doubling filters as the network gets deeper and a final fully connected layer. Uses only one fully connected layer before the final output layer. • Created a text-numeric-categorical functional complex models with multiple inputs, with different modalities, as different data input to classify provider enablement. tf.keras tokenizer is used to tokenise the words then used for generating the Word2Vec embeddings and passing it to keras.Embedding later that fits a LSTM as 1 input layer and 4 hidden layer. One hot encoding is applied on categorical data, followed by feature scaling, then fits an ANN input layer and 3 hidden layer, then connected with keras.layers.concatenate before 2 hidden, 1 output layer.• Developed a text extractor framework which uses docx library and complex logic to extract text, tables and text of the tables from various word documents. This framework/package needs document names and serial number of sections present in documents that needed to generate contracts for providers based on Company’s requirement. It recognises the documents to fetch contract sections from multiple documents to generatare a required contract for providers.• Fabricated the xgboost risk model for assessing the performance of providers. Data used for this project had more than 150 different features of different datatypes. Trained the xgboost model with more than 10 parameters using mlfow. MLflow was used for hyperparameter tunning. -
Data Scientist 1Unitedhealth Group Apr 2020 - Nov 2021• Developed a python version of Atherosclerotic cardiovascular disease (ASCVD) clinal tool/risk predictor, ASCVD generates the risk scores when fed with clinical data, predominantly used by providers to see the risk score for a member. This newly developed python version of ASCVD tool can generate risk score in single go for entire population. Furthermore, applied machine learning techniques and implemented classification models using Sklearn library which ultimately increased the accuracy of predicted risk scores by 10%.• Contributed in building the Python package which is used by team internally. Review and manages GIT’s pull requests for the package.• Created Sqoop scripts to pull down the hive tables on to HDFS. These scripts are being used widely by the team.• Continuously writing Complex SQL queries. Joining multiple tables, sub queries, applying filters, aggregating functions, window functions based on the requirements.• Built more than 20 dashboards using python, plotly, plotly dash, dash components to visualise meaningful insights from the raw data, built a covid dashboard which automatically gets updated every day.• Built couple of web apps comprising of multiple pages using flask, python, html bootstrap. -
Data Science Development AssociateUnitedhealth Group Sep 2018 - Mar 2020Dublin• Developed the text classifier using Long Short term Memory (LSTM). Cleaned the text data, removed stopword, stemming of the texts and turned each text into a sequence of integers/into a vector sequence using Keras tokenizer. Truncated and padded the input sequences so that they are all in the same length for modelling, Keras library were used to implement the model, This architecture efficiently identified the over/under payment for the claims and managed to save big amount.• Built ETL pipelines utilising functionalities of Pyspark and Python. These pipelines were initially built to extract data from different sources, manipulates it then transform it into required format and then finally writes it into hive tables.• Designed and built Tableau dashboards targeting Chronic Kidney Disease(CKD) population. -
InternDucat Oct 2016 - Jan 2017Noida Area, India• Worked on Statistical Package for the Social Sciences (SPSS) software and analyzed the data imported from Excel and SAV files.• Built an SPSS project on what factors predict the likelihood that a drug addicted person would quit the drug or not? (Logistic Regression)• Conducted statistical analysis and implemented data mining techniques.• Created pivot tables and modify spreadsheets to achieve analytical goals.• Performed data manipulation, transformation and cleansing.• Installing, configuring, testing Hadoop ecosystem components.• Learned basic R and implemented classification,prediction models using R.
-
Intern(Undergraduate)Sofcon India Pvt. Ltd May 2015 - Jul 2015Noida, IndiaThis internship was potentially based on Embedded C where I and other two intern worked together and built-up an Automated Car Theft Security system- This embedded system is based on GSM technology. When an unauthorized person tampers a vehicle in which this anti-theft system is settled, the micro-controller commands the GSM module to send a text alert to the vehicle owner that someone is tampering the car.
Siddharth Chaudhary Skills
Siddharth Chaudhary Education Details
-
Amity School Of Engineering2:1 -
Siver Bells Public SchoolScience -
St.Francis School72.57
Frequently Asked Questions about Siddharth Chaudhary
What company does Siddharth Chaudhary work for?
Siddharth Chaudhary works for Unitedhealth Group
What is Siddharth Chaudhary's role at the current company?
Siddharth Chaudhary's current role is Principal Data Scientist.
What schools did Siddharth Chaudhary attend?
Siddharth Chaudhary attended National College Of Ireland, Amity School Of Engineering, Siver Bells Public School, St.francis School.
What skills is Siddharth Chaudhary known for?
Siddharth Chaudhary has skills like C, Data Structure, Core Java, Database Management System, Software Engineering, R, C++, Spss, R Studio, Ssis, Ssas, Power Bi.
Who are Siddharth Chaudhary's colleagues?
Siddharth Chaudhary's colleagues are Patricia Corrigan, Jason Dreier, Bhaskar Kakarla, Daysi Calix, Jim Congleton, Heather Coffey, Michelle Duran Bsn, Rn.
Not the Siddharth Chaudhary you were looking for?
-
Siddharth Chaudhary
Greater Sydney Area -
Siddharth chaudhary
Associate Vp , Uk - Vistas Media Capital | Msc Finance And Investment Banking - Nottingham Business SchoolNottingham -
Siddharth Chaudhary
United States -
3gmail.com, evalueserve.com, bcg.com
-
Free Chrome Extension
Find emails, phones & company data instantly
Aero Online
Your AI prospecting assistant
Select data to include:
0 records × $0.02 per record
Download 750 million emails and 100 million phone numbers
Access emails and phone numbers of over 750 million business users. Instantly download verified profiles using 20+ filters, including location, job title, company, function, and industry.
Start your free trial