Roshan Kumar – Data Scientist & AI/ML Engineer: My Career Journey

My journey in Data Science, Machine Learning and Artificial Intelligence has evolved from learning the fundamentals of programming and data analysis to building and deploying practical AI solutions. I am Roshan Kumar, a Data Scientist and AI/ML Engineer focused on developing intelligent systems that solve real-world problems.

My technical interests include Machine Learning, Generative AI, Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), Computer Vision, NLP, Python and cloud-based AI deployment.

This article shares my career journey, the technologies I have worked with, the lessons I have learned, and the areas of AI and Data Science I continue to explore.

From Data Science to AI/ML Engineering

I started my professional journey by developing a strong foundation in Python, statistics, data analysis and Machine Learning. Working with real datasets helped me understand that successful Data Science is not only about building models but also about understanding business problems, cleaning data, evaluating results and communicating insights.

As my experience grew, my focus expanded toward Artificial Intelligence, Deep Learning, Computer Vision and Generative AI. This progression has allowed me to work across different stages of the AI development lifecycle, from data preparation and experimentation to deployment and production inference.

My Professional Experience

AI/ML Engineer – CMS Computers

In my current role as an AI/ML Engineer, I work on practical Artificial Intelligence and Computer Vision solutions. My responsibilities include working with datasets, supporting model development, validating data, running large-scale inference and helping move AI solutions toward production.

One of the major areas of my work involves Computer Vision for highway inspection. This includes working with object detection systems for identifying and classifying different categories of road defects from survey video data.

This experience has strengthened my understanding of real-world computer vision workflows, including dataset preparation, quality assurance, model inference, large-scale video processing and production-oriented AI systems.

Data Scientist / AI Engineer – Wise Online Stores

At Wise Online Stores, I worked on Data Science and AI projects involving computer vision, predictive analytics, customer segmentation and demand forecasting.

I worked with Python and SQL-based data workflows and explored machine learning approaches for business problems. I also gained practical experience with AWS services, data pipelines and Power BI.

One project involved demand forecasting and customer segmentation, where machine learning helped improve prediction accuracy and support data-driven decision-making.

Data Scientist – Coding Tribes

At Coding Tribes, I worked on machine learning and data analytics problems. My work involved exploratory data analysis, feature engineering, predictive modelling and model optimization.

I also worked with Python and SQL to automate reporting and reduce repetitive manual data-processing activities.

Machine Learning Experience

Machine Learning remains one of the core areas of my technical work. I have worked with both supervised and unsupervised learning techniques and developed projects covering classification, regression, prediction and customer analytics.

My Machine Learning workflow generally includes:

• Understanding the business or technical problem
• Collecting and understanding the data
• Exploratory Data Analysis
• Data cleaning and preprocessing
• Feature engineering
• Model selection
• Training and validation
• Model evaluation
• Deployment and monitoring

I have worked with technologies including Scikit-learn, TensorFlow, Keras, PyTorch, NumPy and Pandas.

Computer Vision and Deep Learning

Computer Vision became an important part of my AI journey as I started working with image and video-based applications.

My experience includes OpenCV, TensorFlow, Keras, CNN-based models, object detection and face recognition.

I have also worked on image-based applications such as food image recognition and nutritional information extraction. These projects helped me understand how computer vision can convert visual information into structured and useful data.

My highway defect detection work further expanded this experience into large-scale video inference and production-oriented Computer Vision systems.

Generative AI, LLMs and RAG

As Generative AI became an important part of modern AI engineering, I expanded my focus toward Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), embeddings and prompt engineering.

I have worked on an LLM-powered RAG chatbot using technologies such as LangChain and FAISS. The workflow involved document processing, chunking, embeddings, retrieval and prompt-based generation.

RAG is particularly useful when an AI application needs to answer questions using a specific knowledge base rather than relying only on information learned during model training.

My current Generative AI interests include:

• Large Language Models
• Retrieval-Augmented Generation
• Vector databases
• Embeddings
• Prompt engineering
• LangChain
• Hugging Face models
• AI-powered applications

Python for Data Science and AI

Python has been one of the most important technologies throughout my career. I use Python across data analysis, Machine Learning, Deep Learning, Computer Vision, APIs and AI application development.

Some of the Python technologies and libraries I have worked with include NumPy, Pandas, Scikit-learn, TensorFlow, Keras, PyTorch, OpenCV, FastAPI and Streamlit.

Python has allowed me to work across the complete journey from data preparation to model development and deployment.

Cloud and AI Deployment

Building a machine learning model is only one part of an AI project. Making that model available through a reliable application is equally important.

I have worked with AWS, Docker, FastAPI, AWS Lambda and GitHub Actions for cloud-based AI deployment and automation.

These technologies have helped me understand how AI solutions can move from experimentation and notebooks toward scalable applications and production workflows.

Some Areas I Have Worked On

• Machine Learning classification and regression
• Customer churn prediction
• House price prediction
• Loan approval prediction
• Customer segmentation
• Demand forecasting
• Computer Vision
• Face recognition
• Food image recognition
• Highway defect detection
• LLM-powered RAG applications
• Data analytics and visualization

Education and Continuous Learning

My academic background is in Computer Science and Engineering. I have also pursued advanced education in Artificial Intelligence and Machine Learning.

However, technology changes rapidly, especially in AI. Because of this, continuous learning has become an important part of my professional journey. I regularly explore new approaches in Machine Learning, Generative AI, Computer Vision, LLMs and AI engineering.

My Technical Skill Set

My current technical areas include:

• Programming: Python, SQL
• Data Science: Pandas, NumPy, data analysis, statistics
• Machine Learning: Scikit-learn, predictive modelling, feature engineering
• Deep Learning: TensorFlow, Keras, PyTorch
• Computer Vision: OpenCV, CNNs, object detection, face recognition
• Generative AI: LLMs, RAG, embeddings, prompt engineering
• AI Frameworks: LangChain, Hugging Face, FAISS
• Deployment: FastAPI, Streamlit, Docker
• Cloud: AWS
• Visualization: Power BI, Matplotlib, Seaborn
• Databases: MySQL, PostgreSQL, MongoDB

What I Have Learned From My Career Journey

One of the biggest lessons from my journey is that practical experience matters as much as theoretical knowledge. Understanding an algorithm is important, but knowing when to use it, how to prepare the data, how to evaluate the result and how to deploy the solution is what makes an AI project useful.

I have also learned the importance of communicating technical concepts clearly. Data Scientists and AI Engineers often work with people from different technical backgrounds, so the ability to explain results and translate business requirements into technical solutions is essential.

Building a Strong Data Science and AI Portfolio

A strong portfolio should demonstrate more than a collection of notebooks. It should show how you approach a problem from beginning to end.

For aspiring Data Scientists and AI/ML Engineers, I recommend building projects that demonstrate:

• Real-world problem understanding
• Data preprocessing and analysis
• Machine Learning model development
• Model evaluation
• Clear documentation
• API or application development
• Deployment where possible
• Business impact and measurable results

This approach helped shape my own learning and professional development.

Explore My Data Science and AI Work

I regularly work on projects and technical topics involving Machine Learning, Generative AI, Computer Vision, Python, LLMs and RAG.

You can explore my technical work, projects and professional information through my online profiles and portfolio.

Frequently Asked Questions

Who is Roshan Kumar?

Roshan Kumar is a Data Scientist and AI/ML Engineer working across Machine Learning, Computer Vision, Generative AI, LLMs, RAG and Python-based AI solutions.

What are Roshan Kumar's main technical areas?

His main technical areas include Machine Learning, Deep Learning, Computer Vision, Generative AI, LLMs, RAG, Python, SQL, AWS and AI deployment.

Does Roshan Kumar work with Generative AI?

Yes. His Generative AI work includes LLM-based applications, RAG systems, embeddings, document processing, prompt engineering and related AI technologies.

What programming language does Roshan Kumar primarily use?

Python is the primary programming language used across his Data Science, Machine Learning, Computer Vision and AI engineering work. SQL is also used for data analysis and data workflows.

What is Roshan Kumar's approach to AI projects?

The approach focuses on understanding the problem, preparing reliable data, developing and evaluating models, and then building practical applications that can be deployed and maintained.

About Roshan Kumar

Roshan Kumar is a Data Scientist and AI/ML Engineer with experience in Machine Learning, Computer Vision, Generative AI, LLMs, RAG, Python, cloud deployment and data analytics.

His work focuses on turning data and AI technologies into practical solutions for real-world problems.

Conclusion

My journey from learning the foundations of Data Science to working on Machine Learning, Computer Vision and Generative AI has been a continuous process of learning and experimentation.

The field of Artificial Intelligence continues to evolve rapidly, and I plan to continue exploring new technologies while building practical AI systems that create measurable value.

If you are interested in Data Science, Machine Learning, Generative AI, Computer Vision, LLMs or RAG, I hope my journey and technical work provide useful ideas for your own learning and projects.