Siddharth Shankar Saraf

Designed and developed LLM-based applications with integrated responsible AI safeguards throughout development lifecycle.

• Reviewed over 50 technical documents and AI use cases, categorizing them per EU AI Act risk levels.
• Developed automated NLP analysis pipeline utilizing sentence transformers to align technical specifications with responsible AI principles.
• Identified 15 high-risk use cases needing compliance enhancements for proactive risk mitigation.
• Collaborated with legal and compliance teams to validate risk assessments and ensure regulatory adherence.
• Delivered more than 20 detailed risk assessment reports, informing AI project approval decisions exceeding $2M.
• Established comprehensive risk severity scoring methodology incorporating technical, ethical, and regulatory dimensions.
• Presented complex technical risks as actionable insights to C-suite executives and board members.

Responsibilities & Key Contributions

• Delivered effective Big data and machine learning solutions and solve complex technical problems.
• Designed and developed Machine Learning/Deep Learning applications for predictive modelling, synthetic data generation, sentiment analysis and trend analysis and deployed in Azure.
• Designing & developing ML applications for one of the largest Multinational bank for- Anti Money Laundering project.
• Finding pattern and trends in data sources and sharing the outlook with seniors and decision makers.
• Used Spark, Databricks Azure to write ML application that leverage big data processing frameworks.
• Find client/user sentiment from the emails they send for products and services offered.
• Used machine Learning Algorithms such as Random Forest, SVM, Regression, XGBoost to predict carbon emission of application servers.
• Involved in different stages of ML pipeline like Data cleaning, EDA, Feature selection, Model creation and Hyperparameter tuning and Model deployment.
• Used Deep Learning algorithms like ANN, CNN and RNN and frameworks like Tensorflow, ScikitLearn to implement Supervised, Unsupervised and Reinforcement Learning.
• Achieved ML accuracy with 92% for one of projects by experimenting with different algorithms.
• Developing, designing ML application for my projects as well as other project as contribution.
• Driving the interaction and partnership between the managers, developers and other stakeholders to ensure active cooperation in identifying as well as defining analytical needs and data needs.
• Automating work flows via Shell Scripting, Apache NiFi and Airflow.
• Monitoring Model deployed and involved in Data Engineering.
• Updated existing ML application to have more metrics and 18% performance increase.
• Deployed model in Google cloud platform built with Tensorflow 2.1.

Responsibilities & Key Contributions

• Designed and developed ETL production pipelines using Hive, PySpark to optimize the extraction of data from raw and disparate sources to clean, integrated analytic datasets used for predictive modeling.
• Designing & developing Big data application for one of the largest Multinational bank for- Anti Money Laundering project.
• Involved in data ingestion from various source systems and processing in HDFS using PySpark.
• Used Spark SQL and Apache Hive to analyze, store data.
• Build ETL pipeline for data to be used for predictive modelling.
• Used machine Learning Algorithms such as Decision Trees, Random Forest, SVM, Regression, K-Means to classify and predict events.
• Involved in different stages of ML pipeline like Data cleaning, EDA, Feature selection, Model creation and Hyperparameter tuning and Model deployment.
• Used Deep Learning algorithms like ANN, CNN and RNN and frameworks like Tensorflow, ScikitLearn to implement Supervised, Unsupervised and Reinforcement Learning.
• Creating Runbooks for application process flow.
• Developing, designing ML application.
• Communicating with different interfacing teams to create data ingestion pipeline.
• Automating work flows via Shell Scripting, CA Autosys and Apache NiFi.
• Monitoring Model deployed and involved in Data Engineering.
• Built different models for regression, classification and clustering and tested them to highest accuracy.
• Deployed model in AWS, used Transfer Learning for applications which had similar problem statement and selected from model zoo.

Responsibilities & Key Contributions

• Designing product which is used by hundreds of different Life Science and Healthcare clients using PySpark, Linux, Machine Learning and Python.
• Developed a machine learning modelling framework in python which showed 20% improvement in sales per customer with an incremental revenue of ~5M.
• Leveraged rule-based extraction learning models to predict likelihood of a customer making a purchase.
• Developing applications and enhancement for proprietary product.
• Migrating existing python (pandas) code into PySpark.
• Processed TBs of data using Spark SQL and Apache Hive.
• Build ETL pipeline from data to be used for predictive modelling.
• Used machine Learning Algorithms such as Decision Trees, Random Forest, SVM, Regression, K-Means to classify and predict events.
• Application development with automation and scheduling.
• Developed application in pyspark, pandas, hive.
• Reporting mechanism of all working scripts and interface connections e.g. Hadoop (Hive)-Tableau interface.
• Involved in building machine learning models, testing them and deploying them in production.
• Integrating with cloud services like AWS, GCP and Azure.

Responsibilities & Key Contributions

• Designing and developing applications which can handle high volume data processing and can integrate with other technologies e.g. Talend, SAP and SQL.
• Developing Big data application for FMCG client.
• Implementing Big data ingestion mechanism with Apache Kafka and scheduling with NiFi.
• Processed TBs of data using Spark SQL and Spark Streaming.
• Used Scala language to write Spark application with RDDs, Spark SQL and Spark Streaming concepts.
• Used Spark Streaming to stream data from external sources using Kafka service.
• Used machine Learning Algorithms such as Decision Trees, Random Forest, SVM, Regression, K-Means to classify and predict events.
• Application development with automation and scheduling.
• Reporting mechanism of all working scripts and interface connections e.g. Hadoop (Hive)-Tableau interface.
• Hortonworks security implementation with Kerberos authentication.

Responsibilities & Key Contributions

• Developing Distributed application which can integrate with other outbound system and can drive insight from the source Data.
• Developing Big data application for Health Care and Retail clients.
• Implementing Big data ingestion mechanism with Sqoop and scheduling with Oozie and Control-M.
• Processed TBs of data using Spark and used Optimization Techniques/Tuning like Kryo serialization and Memory Tuning, etc.
• Used PySpark to write distributed code accessing Spark engine to access data from various source systems and of multiple file types e.g. JSON, TEXT, CSV, PARQUET and process and store in my project.
• Used Spark Streaming to stream data from external sources using Kafka service.
• Used machine Learning Algorithms such as Decision Trees, SVM, Regression, K-Means to classify and predict events.
• Application development with automation and scheduling.
• Reporting mechanism of all working scripts and interface connections e.g. Hadoop-R interface and SAP-HDFS interface.
• Apache Sentry security implementation.

Responsibilities & Key Contributions

• Developing Big Data application for UK Telecom giant and analyzing Network data to predict speed of cables based on loss and other parameters.
• Extracting data from Source systems: RDBMS (Oracle Database) and HIVE (HDFS FileSystem).
• Implementing Big data ingestion mechanism with Sqoop and populating data in my project HDFS.
• Processing TBs of data using Spark and Hive.
• Scheduled various types of jobs with interdependencies using Oozie.
• Used machine Learning Algorithms such as Decision Tree, SVM, Regression, K-Means to classify and predict events.
• Converting Hive based applications to Spark Framework.
• Developed and designed Big data applications to overcome challenge of analyzing large amount of Network data.
• Involved in automating manual processes with Shell scripts and Scheduled them with multiple applications to gain performance improvement both in terms of Time and Cost.