Soumyajyoti Banerjee
Staff Data Scientist at Swiggy l Ex-Amazonian l Ex-Curefit l Ex-Visa
Soumyajyoti Banerjee is a Senior Data Scientist at Swiggy, with 7.87 years of relevant experience. He has previously worked with Amazon, Curefit, and Visa. Soumyajyoti is a skilled technocrat who is proficient in Machine Learning, Natural Language Processing, and Big Data technologies like
Show more
8.6
Years of Experience
Education
iit roorkee indian institute of technology roorkee, iiest shibpur indian institute of engineering science and technology iiest shibpur, uttarpara government high school
Companies
swiggy, swiggy, amazon, cult.fit, visa, paypal, ibm, indian institute of technology roorkee, polaris networks
Reach out to Soumyajyoti Banerjee via Email, InMail and SMS drip
by installing Chrome extension
Soumyajyoti's contact details
Email (Verified)
banXXXXXXXXXXXXXXXXXXXXXXXXXXom
Mobile Number
+91XXXXXXXX34
Github
SouXXXXXXXXXXXXXXee
Experience
2024 - Present
swiggy
Staff Data Scientist
Exploring data science to deliver convenience
2022 - 2024
swiggy
Senior Data Scientist
Currently working on Promise Models. These projects are fuelling better CX and helping to optimise CPD. Delivered multiple project involving predicting leg time of the order cycle. These projects helped to optimize the delivery cycle as well as adhering to promises. These projects involves 1. Big data technologies 2. Regression 3. Deep learning 4. Explainability of the models 5. SOTA Architectures for tabular data (Tab Transformer, FT Transformers & Wide & Deep network) 6. Gen AI (langchain, llm) Experimented with different types of transfer learning technologies, MIMO architecture and embedding techniques to improve the model and extract the business impacts. These architectural changes helped the model to understand the inter dependency of events and impacts at the same time. Technologies used: 1. Pyspark 2. Aws 3. Data bricks As well as helping team members to come up as unit and a significant contributor of the entire eco system.
2019 - 2022
amazon
Data Scientist II
Working on Amazon Finance Automation. Projects deals with NLP, Classification and Clustering. Implementing using Big Data technologies. 1. Developed Automated invoice categorisation framework In this Project created a framework where all the invoices and purchase order will get categorised with pre defined categories. For this problem I have used random forest model for this multiclass classification. As the data is mostly in text form, used HashingTF to convert that to vector. Accuracy : 98.23\% based on invoice count and 96.45% based on spend 2. Developed analysis framework on CBCC Application and GMS penetration in JPCP In this Project created a framework where it will automatically evaluate month on month and YoY growth of CBCC GMS penetration and if any anomaly found it will notify stakeholders regarding that. And Based on requirement anyone can check if the anomaly is coming from any pre defined sets like Prime Non-prime, NTA Non-NTA. Impact: It reduced almost almost 15 man hours monthly to check that if any changes in CBCC GMS penetration in JPCP.
2017 - 2019
cult.fit
Data Scientist
Working on user behaviour and user segmentation based model development. Using time series based models, Neural Networks and Statistical Models. Technology and libraries: Python, SQL, pyFlux, Keras, Tensorflow, theano, Scikit learn, pytorch
2015 - 2017
visa
Senior Software Engineer
working on Spark, Hadoop, Hive, HDFS, Machine Learning, Deep Learning and Big data. Developing a framework which can identify Merchant's name and location using Machine Learning and Natural Language Processing with tools such as Hadoop, Hive, Spark, Pig, Python, Java etc. The framework involved developing a new algorithm for fuzzy string matching with robust phonetic and distance based metrics and incorporating n-gram based filtering to reduce the search space.
2015 - 2015
paypal
RCG 2015
2014 - 2014
ibm
Exterme Blue Summer Intern
Optimize IBM JVM (J9) Garbage Collector (GC) for OSv Ballooning in BlueMix. Objective : Gained knowledge on various garbage collection policies. Objective : Worked on code to use generational garbage collection in BlueMix and OSv. Worked on Generational concurrent garbage collector and developed intelligent way of ballooning to share unused JVM memory among several VMs.
2013 - 2015
indian institute of technology roorkee
Teaching Assistant
1. Tutorial class of Automata Theory. 2. Tutorial class of System programming.
2012 - 2012
polaris networks
SUMMER INTERN
Worked on 4G LTE technology. Worked on testing application of S Gateway Objective : Gained knowledge LTE. Objective : Worked on code of testing application of S Gateway
Experience
81 Skills
algorithms
android
Android
apache
Apache Hive
Apache Spark
Application Programming Interfaces (API)
architecture
architectures
Artificial Intelligence (AI)
Augmented Reality (AR)
Automation
Azure
Big Data
c++
c/c++
Cloud Computing
Collaboration
Data Mining
Data Science
data scientist
Data Scientist
data structures
deep learning
Deep Learning
Design
finance
Flask
forecasting
Git
github
GitHub
Google API
hadoop
Hadoop
Hive
Information Technology
java
Java
kafka
Kafka
keras
Keras
Kernel
LaTeX
LESS
LTE
Machine Learning (ML)
Microsoft Azure
ml
Mobile
mysql
Natural Language Processing (NLP)
neural networks
Neural Networks
NumPy
OOP
open source
operations
optimization
pandas
Parts-of-speech
Polygon
python
Python
Random Forest
Research
Research Scientist
sas
Search
security
social media
Software Engineer
spark
Spring
sql
SQL
tensorflow
test
testing
Windows
Education
2013 - 2015
iit roorkee indian institute of technology roorkee
Master of Technology (M.Tech.)
Computer Science and Engineering
2009 - 2013
iiest shibpur indian institute of engineering science and technology iiest shibpur
Bachelor of Engineering (BE)
Information Technology
1998 - 2009
uttarpara government high school
Higher Secondary
Science
