Soumik Dey profile photo 1 on Tal

Soumik Dey

πŸ‘¨β€πŸ’»MTS-3 @Tessell | Gen AI & Agents | Python β€’ JS | Cloud

Bengaluru

5+ years of work experience

Tessell
Jun 20251 yr 4 mosPresent

Bengaluru, Karnataka, India

NTT DATA
Jan 20241 yr 6 mosJun 2025

Pune District, Maharashtra, India

1. Customer Survey Sentiment Analysis & Product Pain Points πŸ“ Tech Stack: Snowflake (Snowpark & Streamlit) Models: GRU, Random Forest LLM: ML Cortex Functions for Topic Modeling Analyzed customer feedback to understand product pain points. Leveraged Snowflake and advanced models to process sentiment and provide actionable insights. 2. Model Inferencing in Dataiku πŸ“Š Extracted data from PDFs, cleaned and structured it for predictive modeling to forecast product prices. Deployed model metrics on Dataiku Apps for continuous monitoring and optimization. 3. Demand Sensing & Recut Decision for a Fashion Brand πŸ‘— Utilized historical data and trends to optimize demand forecasting for a fast-paced fashion brand. Developed a predictive model for future design SKUs, leveraging Dataiku to integrate various data sources and improve inventory management and production planning. 4. AI-Powered Virtual Technical Support Chatbot πŸ€– Built a Generative AI-based chatbot for automobile manuals and FAQs using LangChain for conversational flow and dynamic document retrieval. Integrated LangSmith to monitor performance and fine-tuned LLMs for domain-specific technical terminology, all with a user-friendly interface built in Streamlit.
Tiger Analytics
Jul 20237 mosJan 2024

Chennai, Tamil Nadu, India

Tiger Analytics
Feb 20221 yr 6 mosJul 2023

On-site

I am part of Data Science Team, building predictive models and optmization tools and automating pipelines, to iteratively producing improved models. 1. Profit Optimization: helping clients to uplift profit by recommending markups for car's parts(Service/Repair,Collision), Providing quartelry recommendations and KPI dashboards on which client can take the adoption accordingly. 2. Extensively worked on web scrapping, explored various libs and tools to extract info from website, invested time to transfrom, manage and standardize the data and also maintaning production ready. 3. Working on Cricket Analytics, where collecting info for all T20 players, trying to understand the perfomance across any condition and trying to build some metrics.(In Discovery Phase) 4. Working on Pharma/Life Science Project, understanding APLD data, and understanding different sources of data which is in different modularity. Scripting Skills: Python, JS, R, SQL Tools : EXCEL, Basic PowerBI Cloud dev: PySpark, Sagemaker MLOps: MLFlow, Docker, KubeFlow Modelling(worked extensively) : Parametric(Linear,Logistic), Non Parametric(DTree, RF), worked on Bayesian modeling, Gradient Descent, Bagging & Boosting Statistics: PMFs,PDFs, Inferential(Hypothesis Testing)
HSBC
Apr 202111 mosFeb 2022

Hyderabad, Telangana, India

Department: Wealth And Personal Banking IT
Indian Institute of Technology, Kanpur
Jun 20203 mosAug 2020

Kanpur, Uttar Pradesh, India

Selected in SURGE Program 2020 under Dr. Preeti Malakar at IITK Subject Of Research- Parallel Computing Github link- https://github.com/soumikopensource/codes

Soumik’s AI Stack

claude code
claude code
cursor
cursor
deepseek
deepseek
gemini
gemini
langgraph
langgraph
langsmith
langsmith

Education

Ramakrishna Mission Vidyalaya

Higher Secondary, PCM

Ramakrishna Mission Vidyalaya

20152 yrs2017
Maulana Abul Kalam Azad University of Technology, West Bengal formerly WBUT

Computer Science and Engineering, Computer Science

Maulana Abul Kalam Azad University of Technology, West Bengal formerly WBUT

20174 yrs2021

Join Soumik and 10k+ other professionals on

Tal Boss

Tal is the app where bosses meet engineers & builders directly, without middlemen

Download Tal Boss

Soumik Dey on Tal

Headline

πŸ‘¨β€πŸ’»MTS-3 @Tessell | Gen AI & Agents | Python β€’ JS | Cloud

About

Who Am I? πŸš€ As a Sr. Data Scientist with expertise in ML/GenAI, MLOps, and Data Engineering, I thrive on driving real-world impact through data-driven solutions that scale. Over the last 5 years, I have contributed to businesses across industries, leading initiatives that accelerated growth, optimized operations, and delivered tangible results. Key Achievements: Improved Forecast Accuracy: Built and deployed predictive models that enhanced demand forecasting by 15%, resulting in optimized resource allocation and increased operational efficiency for a leading fashion brand. Reduced Time-to-Resolution: Developed an AI-powered technical support chatbot with LangChain and RAG, cutting customer query resolution time by 30% for an automobile company. Boosted Model Performance: Optimized models and integrated MLflow for continuous model management, increasing predictive accuracy by 20% in a product pricing project. Streamlined Data Pipelines: Leveraged Snowflake and Dataiku to automate data processing, reducing manual effort by 40% and increasing reporting speed for key stakeholders. Cloud-Native Infrastructure: Architected and deployed scalable solutions on GCP, improving system performance and reducing costs by 25% in cloud resource management. CI/CD Integration: Led MLOps initiatives with Terraform and Grafana, cutting deployment time by 50% and ensuring rapid iteration and delivery in AI and ML workflows. I'm passionate about building end-to-end machine learning systems and collaborating with teams to achieve high-impact outcomes. My work combines technical expertise, a collaborative mindset, and a focus on measurable results. Let's connect if you're looking to drive data-driven success in your organization!

Location

Bengaluru

AI stack

claude code

cursor

deepseek

gemini

langgraph

langsmith

Work experience

Member of Technical Staff 3 at Tessell, Jun 2025 - Present Β· 1 yr 4 mos, Bengaluru, Karnataka, India

Data Scientist at NTT DATA, Jan 2024 - Jun 2025 Β· 1 yr 6 mos, Pune District, Maharashtra, India, 1. Customer Survey Sentiment Analysis & Product Pain Points πŸ“ Tech Stack: Snowflake (Snowpark & Streamlit) Models: GRU, Random Forest LLM: ML Cortex Functions for Topic Modeling Analyzed customer feedback to understand product pain points. Leveraged Snowflake and advanced models to process sentiment and provide actionable insights. 2. Model Inferencing in Dataiku πŸ“Š Extracted data from PDFs, cleaned and structured it for predictive modeling to forecast product prices. Deployed model metrics on Dataiku Apps for continuous monitoring and optimization. 3. Demand Sensing & Recut Decision for a Fashion Brand πŸ‘— Utilized historical data and trends to optimize demand forecasting for a fast-paced fashion brand. Developed a predictive model for future design SKUs, leveraging Dataiku to integrate various data sources and improve inventory management and production planning. 4. AI-Powered Virtual Technical Support Chatbot πŸ€– Built a Generative AI-based chatbot for automobile manuals and FAQs using LangChain for conversational flow and dynamic document retrieval. Integrated LangSmith to monitor performance and fine-tuned LLMs for domain-specific technical terminology, all with a user-friendly interface built in Streamlit.

Senior Analyst - Data Science at Tiger Analytics, Jul 2023 - Jan 2024 Β· 7 mos, Chennai, Tamil Nadu, India

Analyst - Data Science at Tiger Analytics, Feb 2022 - Jul 2023 Β· 1 yr 6 mos, On-site, I am part of Data Science Team, building predictive models and optmization tools and automating pipelines, to iteratively producing improved models. 1. Profit Optimization: helping clients to uplift profit by recommending markups for car's parts(Service/Repair,Collision), Providing quartelry recommendations and KPI dashboards on which client can take the adoption accordingly. 2. Extensively worked on web scrapping, explored various libs and tools to extract info from website, invested time to transfrom, manage and standardize the data and also maintaning production ready. 3. Working on Cricket Analytics, where collecting info for all T20 players, trying to understand the perfomance across any condition and trying to build some metrics.(In Discovery Phase) 4. Working on Pharma/Life Science Project, understanding APLD data, and understanding different sources of data which is in different modularity. Scripting Skills: Python, JS, R, SQL Tools : EXCEL, Basic PowerBI Cloud dev: PySpark, Sagemaker MLOps: MLFlow, Docker, KubeFlow Modelling(worked extensively) : Parametric(Linear,Logistic), Non Parametric(DTree, RF), worked on Bayesian modeling, Gradient Descent, Bagging & Boosting Statistics: PMFs,PDFs, Inferential(Hypothesis Testing)

Software Engineer at HSBC, Apr 2021 - Feb 2022 Β· 11 mos, Hyderabad, Telangana, India, Department: Wealth And Personal Banking IT

Summer Intern at Indian Institute of Technology, Kanpur, Jun 2020 - Aug 2020 Β· 3 mos, Kanpur, Uttar Pradesh, India, Selected in SURGE Program 2020 under Dr. Preeti Malakar at IITK Subject Of Research- Parallel Computing Github link- https://github.com/soumikopensource/codes

Education

Higher Secondary, PCM, Ramakrishna Mission Vidyalaya

Computer Science and Engineering, Computer Science, Maulana Abul Kalam Azad University of Technology, West Bengal formerly WBUT