Skip to content
View mohitkp25's full-sized avatar

Block or report mohitkp25

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
mohitkp25/README.md

Aspiring AI/ML Engineer · M.Tech Artificial Intelligence @ IIT Jodhpur · 14+ years in product & project delivery

I spent 14 years shipping products — running delivery, managing stakeholders, and deciding what was actually worth building. I'm now doing the formal work to build the systems themselves, through an M.Tech in Artificial Intelligence at IIT Jodhpur.

This profile is where that transition happens in public. Expect implementations from first principles, notes on things I'm still figuring out, and steady progress rather than a finished portfolio.

What I'm Building Toward

A career as an applied ML engineer — someone who can take a problem from a vague business need to a model that runs in production, and explain the tradeoffs at every step to people who don't read papers.

Current areas of study:

  • Machine learning foundations — implementing core algorithms from scratch before reaching for libraries
  • Deep learning & neural networks — architectures, training dynamics, optimization
  • Natural language processing — transformers, transfer learning, evaluation methodology
  • GPU computing — CUDA, parallel programming, and where the performance actually comes from
  • Applied ML — the gap between a notebook that works and a system that holds up

What 14 Years Taught Me

The technical skills here are new. These aren't:

  • Choosing the right problem. Most projects fail before a line of code is written. Scoping is a skill.
  • Shipping under constraints. Deadlines, incomplete requirements, and shifting priorities are the normal condition, not the exception.
  • Translating across audiences. Explaining a technical tradeoff to someone who owns the budget but not the codebase.
  • Owning outcomes. Fourteen years of being accountable for whether something actually landed.

I think these compound with ML skills rather than sitting beside them. A lot of models never make it out of a notebook for reasons that have nothing to do with the math.

Working With

Languages: Python ML/Data: NumPy · pandas · scikit-learn · Matplotlib Deep Learning: PyTorch · Hugging Face Transformers GPU: CUDA · Numba Tools: Git · Jupyter · Google Colab

Currently

Completing my M.Tech in Artificial Intelligence at IIT Jodhpur, building out this profile as I go, and open to conversations about applied ML roles where a delivery background is an asset rather than a detour.

Elsewhere

LinkedIn: https://www.linkedin.com/in/mohit-mathur-project-manager/ Microsoft: https://learn.microsoft.com/en-us/users/mohit-mathur-project-manager/ Google: https://me.developers.google.com/u/mohit-mathur

Popular repositories Loading

  1. K-Fold-Cross-Validation K-Fold-Cross-Validation Public

    The primary purpose of this assignment was to code a K-Nearest Neighbors (k -NN) classifier in Python with the iris data, and subsequently perform our own K-Fold Cross-Validation in Python to under…

    Python

  2. Decision-Trees Decision-Trees Public

    - Implement a Decision Tree from scratch using Gini impurity and Entropy as splitting criteria. - Understand the recursive partitioning process used to construct decision trees. - Explore the role …

    Python

  3. matrix-exponentiation matrix-exponentiation Public

    CUDA-based Matrix Exponentiation using Exponentiation by Squaring. Implements CPU and GPU versions in Python/Numba and compares performance for large matrix powers (A^100).

    Python

  4. Sanskrit_to_English-Neural-Machine-Translation Sanskrit_to_English-Neural-Machine-Translation Public

    Fine-tuning IndicTrans2 for Sanskrit-to-English translation on a low-resource parallel corpus, with BLEU/BERTScore evaluation and error analysis

    Jupyter Notebook

  5. Random-Forest Random-Forest Public

    Implementation of the Random Forest Algorithm

    Python

  6. linear-regression linear-regression Public

    Linear Regression Using Gradient Descent

    Python