About Me

I build machine learning systems that transform raw data into actionable predictions

View Resume

The Beginning

My journey into machine learning started with a simple question: "Can data actually predict what happens next?"That curiosity led me from writing my first Python script to building end-to-end ML pipelines.

Currently pursuing B.Tech in Computer Science (Data Science) at National Institute of Science and Technology, I've complemented my formal education with hands-on projects and professional certifications from the University of Michigan and Coursera.

What started as curiosity has become a passion for turning messy real-world data into actionable insights that solve real problems.

How I Think Now

After building several ML systems, my engineering philosophy has crystallized around a few core beliefs:

1.
Data quality beats model complexity.

A simple model on clean data outperforms a complex model on messy data.

2.
Understand before you optimize.

The best predictions come from understanding why something happens.

3.
Reproducibility is non-negotiable.

If you can't reproduce your results, you don't understand them.

Education

B.Tech Computer Science & Engineering

Data Science

In Progress

National Institute of Science and Technology

August 2024 – August 2028

Higher Secondary (XII Science)

CBSE

Completed

St. Xavier's High School

April 2020 – May 2022

Professional Courses

Applied Machine Learning in Python

University of MichiganCoursera

April 2024 – June 2024

  • Supervised & unsupervised learning algorithms
  • Model training, testing, and evaluation
  • Feature engineering & data preprocessing

Machine Learning with Python

CourseraCoursera

January 2024 – March 2024

  • Supervised & unsupervised learning algorithms
  • Model training, testing, and evaluation
  • Feature engineering & data preprocessing

Values

Intellectual Honesty

I say "I don't know" when I don't. I share failures alongside successes.

Continuous Learning

ML evolves fast. I dedicate time weekly to learning new techniques.

Clear Communication

Complex ideas should be explained simply. Jargon hides weak thinking.

Ship & Iterate

Perfect is the enemy of good. Working solutions beat theoretical perfection.

Languages

OdiaNative
HindiFluent
EnglishProfessional

Working Style

  • I prefer async communication for deep work, with sync catchups for alignment
  • I document as I build — code comments, READMEs, and decision logs
  • I ask questions early rather than making assumptions
  • I value honest feedback and give it respectfully in return

Environments I Avoid

Being upfront about what doesn't work for me:

  • Teams where metrics don't matter (decisions by opinion, not evidence)
  • Environments that punish honest mistake acknowledgment
  • Pressure to deploy models without proper validation

Skills Overview

Loading chart...

What People Say

Feedback from people I've worked with and learned from

"Kumlesh demonstrated exceptional problem-solving skills during our ML training program. His ability to break down complex problems into manageable components is impressive."
TM

Training Mentor

Senior Data Scientist at Analytics & ML Training Program

Mentor
"A dedicated learner who consistently delivers quality work. His churn prediction project showed real understanding of both the technical and business aspects of ML."
PS

Project Supervisor

ML Team Lead at AI Solutions Lab

Supervisor
"Great attention to detail in data preprocessing and feature engineering. Always eager to learn and improve. Would recommend for any ML/data science role."
PDS

Prof. Data Science

Associate Professor at NIST

Professor