``

Mohit Kumar.

|

AI Architect with over 4.5 years of experience in delivering scalable data-driven solutions into production.

About Me

Mohit Kumar

Hello! I'm Mohit Kumar, an AI Architect with over 4.5 years of experience in delivering scalable data-driven solutions into production. I specialize in Machine Learning, Data Science, Deep Learning, Computer Vision, NLP (LLMs), and MLOps & DevOps.

I intend to be a part of an organization where I can constantly develop my skills and use them to the best of my ability for the organization's growth. With a strong foundation in both technical and practical aspects of ML engineering, I approach each project with creativity and technical precision.

4.5+ Years Experience
15+ Projects
100000+ Lines of Code

Education

Master of Science, Information Security

Indira Gandhi National Open University

July 2022 – July 2024

Bachelor of Technology, Electronics & Communication Engineering

Delhi Technical Campus

Aug 2017 – May 2021

Experience

Machine Learning & MLOps Engineer

Sirion Dec 2023 - Present
  • Collaborating on developing scalable enterprise-grade CLM software with product teams and architects.
  • Currently acting as bridge between ML and DevOps teams for cloud infrastructure migration across platforms.
  • Leading the service migration from GCP Kubernetes to AWS EKS for scalable ML infrastructure.
  • Implementing zero-scale kubernetes clusters using KEDA to minimize off-peak infrastructure costs.
Machine Learning MLOps Kubernetes AWS

Senior Machine Learning Engineer

Aligne Consulting Aug 2021 - Dec 2023
  • Designed and implemented robust and highly scalable machine learning microservices for Axion (SaaS Product).
  • Executed POCs using Large Language Models (LLM) for the upcoming ML use cases in the product.
  • Actively contributed to the migration & enhancement of on-premises solutions to the cloud-based SaaS environment.
  • Ensured the smooth operation and continuous improvement of the product infrastructure by maintaining and enhancing the CI/CD pipelines, ensuring optimal availability and performance.
Machine Learning LLMs Microservices CI/CD

Data Scientist

Sankshit Private Limited: AAIENA Sep 2020 - July 2021
  • Redesigned and modernized the user's apparel shopping experience using Deep Learning and Computer Vision.
  • Led the team of 5 developers to launch Sizing product which accurately returns the body measurements.
  • Worked on Ettire that assists users to view their virtual version (3D) draped in different clothes.
  • Contributed to deployment & scalability of products for approx. 100000 users footfall every day.
Deep Learning Computer Vision 3D Modeling Scalability

Skills

Projects

Scalable spaCy Server

Scalable spaCy Server

The project aimed at delivering a scalable solution for sensitive data recognition in enterprise data. Application also has the capability of custom model training and in-memory loading in python multiprocess. The spaCy large model was used and served through FastAPI using LFU Cache.

Server withstood 3000 requests in 5 minutes in performance test with an average response time of 800ms.

Python spaCy FastAPI Caching
Image Super Resolution

Image Super Resolution

The project aimed at obtaining high-resolution images from low-resolution images using end-to-end mapping. Implemented Image Super-Resolution Using Deep Convolutional Networks research paper (2014).

Designed SRCNN architecture with Feature extractor, non-linear mapping, reconstruction as components. Evaluated model results using PSNR, SSIM, and MSE as image quality metrics & obtained precise results.

Deep Learning Computer Vision CNN Image Processing
Textual Similarity Analysis

Textual Similarity Analysis

The project aimed at delivering a solution for detecting semantic similarity in enterprise data. Universal Sentence Encoder was used and served through TensorFlow Serving for text embeddings.

The cosine similarity metric was used for the text similarity search between the embedding vectors. Application withstood 6000 requests in 5 minutes in performance test with an average response time of 160ms.

NLP TensorFlow Embeddings Semantic Search

My Medium Articles

I share my technical knowledge and insights on Medium. Check out some of my popular articles:

LLama 3 LLM Article

Running LLama 3 LLM with vLLM Library at Scale

Learn how to efficiently deploy and scale Large Language Models using the vLLM library.

June 30, 2024
Read Article →
TensorFlow Serving Article

Deploying a TensorFlow Model with TensorFlow Serving and Docker

A step-by-step guide to deploying machine learning models in a production environment.

January 15, 2023
Read Article →

Get In Touch

Let's Connect

I'm always open to discussing new projects, creative ideas, or opportunities to be part of your vision. Feel free to reach out through any of the channels below.

Email

krmohit101@gmail.com

Phone

Available upon request

LinkedIn

mohitkumar1999

Twitter/X

@imohit_kr