Shounak K.

Lead ML Engineer

Bengaluru, Karnataka, India13 yrs 8 mos experience

Highly StableAI Enabled

Key Highlights

Expert in Natural Language Processing and Deep Learning.
Proven track record in predictive modeling and data visualization.
Holds a U.S. patent and published research in applied deep learning.

Stackforce AI infers this person is a Data Scientist with expertise in SaaS and Ad Tech industries.

Contact

Skills

Core Skills

Machine LearningDeep LearningNatural Language Processing (nlp)Data SciencePredictive ModelingJ2ee Web Services

Other Skills

Ad TechAlgorithmsAnalytical SkillsApache KafkaApache SparkArtificial Intelligence (AI)Azure DatabricksBERTCC++Cloud ComputingConvolutional Neural Networks (CNN)Core JavaData CleaningData Mining

About

As a data scientist at InMobi, I apply my natural language processing and transformer skills to various tasks, such as named entity recognition, named entity disambiguation, and news article classification. I use state-of-the-art models, such as XLNet,BERT to achieve high accuracy and performance. I also leverage my experience in Spark and Superset to create a dashboard for monitoring AB testing experiments, and to predict click through rate and bid price using gradient-boosted trees. Used deep learning models for various prediction tasks. Finetuned BERT based model to generate user embedding and use in user targeting model. I have one approved U.S. patent and one publication in applied deep learning, demonstrating my innovation and research capabilities. I have over 9 years of experience in software engineering, machine learning, and data science, and I hold a M.Tech in Computer Science from IIT Hyderabad and a PG Diploma in Data Analytics from IIIT Bangalore. My goal is to solve complex and impactful problems using data and AI.

Experience

Salesforce

Lead Member of Technical Staff

Aug 2025 – Present · 7 mos · Bengaluru, Karnataka, India · On-site

Joining Agentforce as Lead ML/AI Engineer

Large Language Models (LLM)Vector DatabasesRetrieval-Augmented Generation (RAG)Deep LearningMachine LearningPredictive Modeling

Inmobi

Data and Applied Scientist - III

Mar 2022 – Jul 2025 · 3 yrs 4 mos · Bengaluru, Karnataka, India · On-site

Named Entity Recognition (NER) and Named Entity Disambiguation (NED) on news articles.
Using Transformers (XLNet) to predict the correct news article class.
Created a dashboard using Apache Spark and Superset (for visualization) for monitoring AB testing experiments.
Click Through Rate prediction using traditional gradient-boosted trees (LightGBM).
Bid price prediction.
Distributed Deep Learning training of more than 1 billion data points using Ray on AKS (Azure Kubernetes Service) which reduced the training cost by 50% over Horovod on Databricks.
Finetuned Large Language Models (BERT, T5) and created user embeddings from them to use in other Tensorflow based deep learning models , which improved precision@5% recall by ~20%

BERTPredictive ModelingApache SparkPyTorchAnalytical SkillsDeep Learning+19

Buddi ai

Research Associate

Jan 2021 – Feb 2022 · 1 yr 1 mo · Chennai, Tamil Nadu, India

Working on Data Preprocessing to Deep Learning model building entire pipeline for the medical code prediction task.
Working on Transformer based model (GPT2) on medical data to generate synthetic medical text from only keywords.

BERTPredictive ModelingApache SparkPyTorchAnalytical SkillsDeep Learning+17

Career break

Professional development

Jan 2018 – Dec 2020 · 2 yrs 11 mos · Hyderabad, Telangana

Took break to pursue higher studies (MTech) from IIT Hyderabad

Oracle india pvt. ltd

Senior Applications Engineer

Jun 2016 – Jan 2018 · 1 yr 7 mos · Hyderabad Area, India

Worked as a Product development engineer at Oracle Fusion HCM team. I worked as a developer in the U.S Payroll localization team.
Worked on the various critical bug fixes in the product.
Worked on the development of the tax card calculation functionality.

J2EE Web ServicesStrutsSQL

Tata consultancy services

System Engineer

Mar 2012 – Apr 2016 · 4 yrs 1 mo · Kolkata Area, India

Was involved in the development of the TCS MasterCraft JavaProfiler solution.
This tool uses the Java Byte Code Instrumentation technique to profile any Java application.
Had implemented new features in the tool like Java thread dump analysis, Java heap dump analysis, collecting exception and unexited method details, Garbage collection details .
Also , as a part of my work, I had filed 1 patent, which has been approved ( https://patents.google.com/patent/US9990273B2/en )
Was involved in the development of the TCS MasterCraft Application Performance Monitor.
This tool monitors various aspects of an application, like :
monitoring database health, application server response time, disk usage of app server and DB server, the memory usage of servers, pinging hosts, etc.
Worked as a lead developer. Apart from my work, I used to mentor fresher who were coming to my team.

J2EE Web ServicesStrutsTomcatjava swing