SAI KRISHNA KISHORE BEATHANABHOTLA
+1 - |-| https://www.linkedin.com/in/saikrishnakishore
•
•
•
•
•
•
•
•
EXPERIENCE SUMMARY
8+ years of overall experience working with diverse technology stack related to Data Engineering and Machine Learning.
With 5+ years of strong development experience as a Machine Learning Engineer and a Data Scientist, developed
applications based on Natural Language Processing and recommender systems.
Architected, designed and developed several Active Learning systems which can perform online learning and predictions.
Experience working with attention mechanisms, transformer architectures and pre-trained models like BERT and various
neural network architectures using RNN, Seq2Seq, Transformers, Attention, and Memory networks.
Designed and Developed systems to explain the rationale of ML models using Explainable Neural Networks.
Good hands-on experience working with tools like Neptune.ai, Tensorboard and Streamlit.
Good understanding about various Neural Network architectures like GAN, SSD, VGG, ResNet and Reinforcement
Learning algorithms like Q-Learning, DQN, DDQN, Policy Gradient, Actor Critic, and A3C.
Hands-on experience working with AWS.
Programming Languages
Deep Learning Frameworks
Machine Learning Frameworks
NLP Packages/Frameworks
Distributed Processing Frameworks
Data Munging and manipulation
Hadoop Ecosystem
Version Control and Deployment
Workflow and Schedulers
Databases
TECHNICAL STRENGTH
Java, Scala, Python
Keras, Tensorflow, PyTorch
PySpark, Scikit-learn, Gym
BERT, Gensim, Glove, NLTK
Spark, Flink
Pandas, Numpy
HDFS, Hive, Impala, Kafka
Git, Jenkins, Docker, Artifacts, Kubernetes
Oozie, Airflow
Neo4J, HBase, MongoDB, Greenplum, Teradata, MySQL
WORK EXPERIENCE
Collective[I]
Role – Machine Learning Engineer/Big Data Engineer
Ecosystem: Tensorflow, Scikit-learn, BentoML, Pandas, PySpark, AirFlow, Hive, MongoDB, HBase
Roles and responsibilities:
•
•
Montreal, Canada
Mar 2020 – till date
Architected and developed Active learning identity resolution systems to unify/de-duplicate the data across the board.
Architected, designed and implemented Graph based recommender system to generate online recommendations for various use
cases like best buyer to reach, best person to contact etc.
• Architected and developed various scikit-learn based machine learning modules for forecasting revenue and win/loss of a sales
opportunity using techniques like time-series forecasting, boosting, bagging and ensemble models.
• Improved performance of Spark applications to consume lesser resources and time to process the data.
• Developed multiple Spark applications responsible for data ingestion into systems like Neo4J.
• Developed AKKA based ingestion techniques to ingest data into Neo4J.
Innominds Software
Project – AI Feed
Hyderabad, India
Role - Machine Learning Engineer and Data Scientist
Apr 2016 – Mar 2020
Ecosystem: PySpark, Keras, Tensorflow, Hive, Tensorboard, TensorFlow Serving, W&B Artifacts, Neptune.ai, Docker
Roles and Responsibilities:
• Was holding dual responsibilities as a Machine Learning Engineer and as a Data Scientist.
• Designed and developed a scalable news article recommender system in iterations by initially building the system for sessionbased news article recommendation and later evolving it to include user’s history based on past sessions.
• Developed attention models to track user’s short term and long term interests to generate custom home pages.
• Developed custom metrics to evaluate performance of models and benchmarked the performance over time.
Teradata India
Product - Teradata Listener
Hyderabad, India
Role - Big Data Engineer
Nov 2014 - Apr 2016
Ecosystem: Spark, Kafka, Mesos, Marathon, Docker.
Roles and Responsibilities:
• Designed and developed a high-volume streaming ELT system which can consume the data from high volume sources.
• Implemented custom Spark streaming block management to circumvent the overhead of uneven blocks.
• Designed and developed RESTful endpoints to ingest data into Kafka, Docker was used to containerize the services.
• Implemented Spark Streaming to consume the data from Kafka and parse the data to the destination schema format and push
the data.
• Implemented custom Spark streaming block management to get rid of overheads of uneven blocks.
• Responsible for benchmarking throughput and latency of the data ingestion module and data parsing modules.
• Designed and developed APIs to constantly push metrics to ElasticSearch and monitor using Kibana.
Valuelabs LLP
Project - TAG
Hyderabad, India
Role - Big Data Engineer
Jan 2013 - Nov 2014
Ecosystem: MapReduce, Java, Impala, Hive, ElasticSearch.
Roles and Responsibilities:
• Develop framework to ingest data from Teradata to hadoop based ecosystem using MapReduce.
• Implemented custom compression format to encrypt data using AES-256 encryption and then compress the data using
LZO compression.
• Developed custom Hive serde to read the data from the encrypted data blocks.
• Expose the migrated data using Hive and Impala.
• Developed canned reports using HOLAP model to generate reports on top of migrated data.
• Developed dataflow pipelines using rule engines configurable by business use cases using Activiti BPMN.
CERTIFICATIONS
•
•
•
•
•
AWS certified solutions architect – Associate.
Databricks certified Spark developer.
Cloudera Certified Developer for Apache Hadoop (CCDH).
Oracle Certified Associate for Java 7 (OCJA).
Oracle Certified Professional for Java 7 (OCJP).
•
•
M.Tech - Software Systems from BITS Pilani, specialization in Data Analytics from 2016 - 2018, with CGPA of 8.8.
B.Tech from JNTU Hyderabad, specialization in Electronics Engineering from 2008 - 2012, with an aggregate of 77%.
EDUCATION
