Sourav Kumar

AI/ML Engineer
Near SAP labs, Whitefield, Kundalahali, 560037, Bangalore North, India.

About

Highly analytical AI/ML Engineer with 1+ years of experience driving data-driven innovation and process optimization across diverse industries. Proficient in Python, SQL, and advanced ML/NLP techniques, I specialize in developing scalable models and deploying solutions that translate complex data into actionable insights. Proven ability to deliver significant business impact, including reducing manual effort by 25% and boosting operational efficiency by 18%.

Work

Rare Metals and Chemicals
|

Junior Data Analyst

Ranchi, Jharkhand, India

Summary

As a Junior Data Analyst, I automated reporting, designed interactive dashboards, and collaborated cross-functionally to enhance operational efficiency and quality.

Highlights

Automated daily and monthly production reporting using Python and Excel macros, reducing manual effort by 25% and streamlining data processing workflows.

Designed and deployed interactive dashboards in Excel and Power BI, providing real-time insights into key performance indicators (KPIs) such as yield, material consumption, and batch-wise flow.

Collaborated cross-functionally with process and production teams to identify and resolve operational inefficiencies, implementing data-backed improvements that enhanced output and product quality.

Learn Vista Pvt Ltd
|

Data Science Intern

Remote

Summary

As a Data Science Intern, I developed and deployed ML models, applied advanced NLP techniques, and designed Power BI dashboards to support data-driven decision-making.

Highlights

Developed and deployed robust ML models for critical applications including fraud detection, healthcare prediction, and recommendation systems, utilizing Python, Scikit-learn, and Pandas.

Utilized advanced NLP techniques and generative AI tools (ChatGPT, Gemini) for sentiment analysis, keyword extraction, and automated content generation from complex datasets.

Created dynamic Power BI dashboards and visualizations with Matplotlib and Seaborn, translating complex data trends into actionable insights for strategic decision-making.

Volunteer

Entrepreneur Club (I.K. Gujral Punjab Technical University)
|

Secretary

Kapurthala, Punjab, India

Summary

Led planning and execution of innovation events and workshops, fostering entrepreneurial spirit and engagement within the university community.

Highlights

Led planning and execution of 3+ innovation events and workshops focused on entrepreneurship, coordinating with faculty, industry mentors, and over 100 students to ensure high engagement.

Education

I.K. Gujral Punjab Technical University (Main Campus)
Kapurthala, Punjab, India

B.Tech

Computer Science and Engineering

Grade: 75%

Languages

English

Certificates

Microsoft Azure Fundamentals

Issued By

Microsoft

IBM Advanced Data Science & AI Specialization

Issued By

IBM

Tableau & Power BI Certification

Issued By

Learnbay

Skills

Programming & Databases

Python, SQL, MySQL, MS SQL Server, MongoDB, ChromaDB.

Core AI & Machine Learning

Supervised Learning, Unsupervised Learning, Deep Learning, NLP (BERT), CV (CNN, OpenCV), GANs, Autoencoders, EDA Feature Engineering, A/B Testing, SHAP LIME, Time Series (ARIMA, LSTM).

Libraries & Frameworks

Scikit-learn, TensorFlow, Hugging Face, Transformers, Pandas, NumPy, Matplotlib, Seaborn.

MLOps & Deployment

Streamlit, FastAPI, MLflow, Docker, Kubernetes, Git, DVC, Gradio, Dialogflow.

Cloud & Generative AI

AWS, Azure, GCP, LangChain, Dialogflow, ChatGPT, Gemini, OpenAI API.

Projects

Conversational RAG Chatbot (Gemini + LangChain)

Summary

Developed a session-aware Generative AI chatbot leveraging RAG architecture for real-time information retrieval and interactive web lookups.

Emergency Image Classification using Deep Learning

Summary

Built an end-to-end deep learning model for binary image classification, specifically for detecting emergency versus non-emergency vehicle scenes.

NLP Sentiment Analysis – Political Speech

Summary

Engineered a comprehensive NLP pipeline for sentiment classification on political speeches, including text preprocessing, tokenization, and visualization.