Skip to content
View DevSharma03's full-sized avatar

Block or report DevSharma03

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 supported. This note will be visible to only you.
Report abuse

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

Report abuse
DevSharma03/README.md

Hi, I'm Devashish Sharma 👋

AI/ML Engineer • Data Scientist • Full-Stack Developer

I build scalable, production-ready AI systems using Machine Learning, NLP, LLMs, and Cloud technologies.


👋 About Me

I’m an AI/ML Engineer & Data Scientist focused on building production-ready, real-world AI systems — from research and experimentation to deployment and scale.

  • 🧠 Hands-on experience with LLMs, RAG systems, NLP, traditional machine learning, deep learning, and neural networks
  • ☁️ Experience deploying and scaling ML solutions on AWS & GCP, including stateful and stateless architectures
  • ⚙️ Strong expertise in end-to-end ML pipelines (data ingestion → model training → APIs → UI)
  • 🛡️ Hands-on experience in full-stack development, enabling seamless integration of AI into real products

Currently pursuing a B.Tech in Artificial Intelligence & Data Science and actively working on impact-driven, production-grade AI platforms.


🧠 Core Expertise (2025 Focus)

  • LLMs & Agentic AI: Retrieval-Augmented Generation (RAG), prompt engineering, multi-agent and orchestrator–worker system architectures

  • Machine Learning: Supervised and unsupervised learning — classification, clustering, regression, feature engineering, and model evaluation

  • Natural Language Processing (NLP): Text preprocessing and representation, tokenization (WordPiece/BPE), embeddings, transformer-based models (BERT, RoBERTa, DistilBERT), text classification, semantic similarity, and information extraction (Foundational NLP: lemmatization, stemming, stop-word handling, TF-IDF)

  • MLOps & Deployment: FastAPI-based inference services, Dockerized ML workflows, CI/CD pipelines, scalable and production-ready model deployment

  • Cloud Platforms: AWS (EC2, Lambda), Google Cloud Platform (GCP)

  • Full-Stack Development: React, Node.js, Express.js, MongoDB, Mongoose, SQL, GraphQL — building end-to-end AI-powered platforms


🧰 Tech Stack

Languages:
Python · Java · JavaScript · SQL · R

AI / ML:
Scikit-learn · TensorFlow · PyTorch · spaCy · NLTK

LLMs & GenAI:
RAG · LangChain · LangGraph · Prompt Engineering

Backend & APIs:
FastAPI · Node.js · Express · Django · RestAPI

Frontend:
React.js · Bootstrap · Tailwind CSS · EJS Templates

Data Engineering & Pipelines:
Pandas · NumPy · Feature Engineering · ETL · EDA · Streaming Basics

Databases:
MongoDB · PostgreSQL · MySQL

Cloud & DevOps:
AWS · GCP · Docker · GitHub Actions · CI/CD


🧪 Featured Projects

🔹 LifeSync — AI-Powered Productivity & Life Analytics Platform

Tech Stack: Python, spaCy, LightGBM, GaussianMixture, LLMs (Groq), FastAPI, React.js, Express.js, MongoDB, Prompt Engineering

  • Built a full-stack platform to help users improve their health, finances, productivity, & well-being by tracking habits, goals & utilities.
  • Integrated a RAG-based AI assistant powered by LLMs, leveraging an orchestrator-worker architecture to deliver scalable, personalized, data-driven insights.
  • Developed ML models for score generation and user clustering with 95%+ accuracy and used synthetic data for training with real time data to enhance model performance and scalability.

🔹 ApplyXpert — ATS Resume Analyzer (NLP)

Tech Stack: Python, spaCy, scikit-learn, TF-IDF, React.js, Node.js, Express.js, MongoDB, Fast-API

  • Developed an end-to-end NLP-based AI-powered ATS resume analyzer with interactive UI to compare resumes against job desc.
  • Implemented personalized, AI-driven feedback to improve CV quality, achieving 97%+ accuracy in CV–job description alignment and enabling more effective job applications.

🔹 Network Traffic Anomaly Detection (Cybersecurity)

Tech Stack: Python, R, Fast-API, Isolation Forest, Leaflet, HTML

  • Designed and applied an end-to-end network traffic anomaly detection system using Isolation Forest, enabling identification of suspicious patterns and security threats with alert generation model.
  • Deployed model via Fast API with real-time Leaflet dashboard for visual analysis, achieving 94% detection accuracy and supporting faster threat identification and data driven security decision.

🔹 Stealth Lateral Movement Detection (GNN – Research)

Tech: Python · Graph Neural Networks · Docker · AWS

  • Modeled enterprise networks using graphs for advanced threat detection
  • Improved detection accuracy by ~40%
  • Deployed scalable inference using FastAPI + Docker

🎯 Currently Exploring

  • Agentic AI & LLM automation
  • AI-driven cybersecurity systems
  • AI-driven Agricultural Systems
  • Advanced MLOps & scalable model serving
  • Generative AI + real-time analytics

🤝 Let’s Connect

💬 Open to AI/ML, Data Science, SDE, and Research opportunities
📩 Feel free to reach out via LinkedIn or Email

⭐ If you like my work, consider following or starring my repositories!

Pinned Loading

  1. ApplyXpert ApplyXpert Public

    This application helps job seekers improve their resumes by analyzing them against job descriptions using NLP techniques.

    JavaScript 3

  2. LifeSync LifeSync Public

    LifeSync is a multi-system repository that brings together a Backend, Frontend, LLM service, and Model artifacts to provide an integrated personal assistant / life-management platform. This README …

    JavaScript 2

  3. ERP-Website ERP-Website Public

    A fully functional Enterprise Resource Planning (ERP) website designed to manage business processes efficiently. This system integrates multiple modules, providing a seamless user experience for ha…

    EJS 3

  4. Credit_Score_Prediction_Model Credit_Score_Prediction_Model Public

    The Credit Score Prediction Model is a machine learning–based solution designed to predict an individual’s credit score using financial and demographic attributes. This project demonstrates a compl…

    Jupyter Notebook 1

  5. Network-Traffic-Detection-Model Network-Traffic-Detection-Model Public

    A smart anomaly detection system built using R for intelligent network traffic analysis and Python Flask for web deployment. This model identifies unusual traffic patterns to flag potential fake or…

    Jupyter Notebook 3

  6. Spam_Mail_Detection_Pipeline Spam_Mail_Detection_Pipeline Public

    Jupyter Notebook 1