{ ... }
=>

Hi, I'm Mayank Bansal

I build

Specialized in AI, Machine Learning, and Python Development

View My Work Contact Me
+
✦
✦
</>
// ai
::
ABOUT ME

Engineering Intelligent Systems

A glimpse into my background, engineering philosophy, and passion for building AI.

Mayank Bansal

Who am I?

I'm an AI/ML engineer in training and a 3rd-year BTech CSE student focused on building and shipping intelligent systems, not just experimenting with models.

My work spans RAG, Agentic AI, Deep Learning, and NLP, with hands-on experience building retrieval pipelines, LLM-powered applications, neural network and computer vision systems, and production-oriented APIs. I enjoy working across the stack, from data and model development to retrieval, orchestration, evaluation, and deployment.

I learn by building, validate through benchmarks and real-world testing, and continuously push my projects from working prototypes toward robust, useful systems.

Education

BTech in Computer Science & Engineering

Lovely Professional University — 3rd Year

CGPA: 9.41 / 10
Download CV View CV
+
✦
✦
{ }
=>
CREDENTIALS

Verified Certificates

Continuous learning in Artificial Intelligence, Machine Learning, and Software Architecture.

Certificate 1
+
✦
✦
✦
{ }
01
// stack
SKILLS

Technologies I Work With

A curated set of tools and technologies I use to build, experiment, and ship AI systems.

Programming Languages

Languages I use for development and problem solving.

Python
C++
C
Java
JavaScript
SQL
HTML/CSS

RAG & Agentic AI

Tools for building LLM applications, retrieval systems and agentic workflows.

LangChain
LangGraph
RAG
FAISS
ChromaDB
Gemini API
Sentence Transformers
Fine Tuning

Machine Learning & Deep Learning

Frameworks and techniques for building and training models.

PyTorch
TensorFlow
Scikit-learn
Deep Learning
Neural Networks
CNNs
Computer Vision
NLP
Model Evaluation

Backend & Deployment

Building and deploying scalable applications.

FastAPI
Flask
REST APIs
Deployment
UI Design

Tools & MLOps

Tools for development, experiment tracking and deployment.

Git & GitHub
MLFlow
Docker
MLOps
Jupyter Notebook
VS Code
CONTINUOUSLY LEARNING BUILDING IMPROVING
+
✦
✦
✦
✦
</>
[ ]
&&
PORTFOLIO

Featured AI Projects

Autonomous agents, RAG pipelines, and computer vision systems built for real-world impact.

AI Career Copilot

AI Career Copilot

  • Architected an agentic RAG career assistant and ATS scorer with LangGraph state machine orchestration and human-in-the-loop resume diff review.
  • Engineered a dual-provider LLM pipeline (Groq cloud & Ollama local) with ChromaDB semantic retrieval, achieving 90% Recall@4, 0.83 MRR@4, and 0.81s P50 latency.
  • Containerized production microservices stack with Docker and exposed modular endpoints via FastAPI.
Python FastAPI LangChain LangGraph RAG ChromaDB Groq Ollama Docker
GitHub
Technical Debug Assistant

Technical Debug Assistant

  • Engineered a production-style Multi-Hop RAG debugging assistant using FAISS vector indexing and Groq-hosted Llama 3.3 70B for context-grounded diagnostics.
  • Built an end-to-end ingestion pipeline with metadata-aware chunking, semantic vector embeddings, and multi-document context synthesis.
  • Integrated dynamic document uploads for live knowledge base updates, transparent source attribution, and retrieval score visualization.
Retrieval-Augmented Generation (RAG) Python FAISS Groq API Llama 3.3 70B Semantic Search Prompt Engineering
GitHub
PDF Question Answering System

PDF Question Answering System (RAG)

  • Built an end-to-end PDF question answering pipeline using Sentence Transformers embeddings and a FAISS vector index for low-latency similarity search.
  • Integrated Gemini API with structured prompt engineering and safety guardrails to enforce context grounding and mitigate hallucinations.
  • Developed an interactive web interface with Flask for dynamic PDF ingestion, chunking, and document querying.
Python Flask FAISS Sentence Transformers Gemini HTML CSS JavaScript
GitHub
Computer Vision Image Search

Visual Search Engine

  • Implemented an end-to-end visual similarity search engine using a pre-trained ResNet50 CNN in PyTorch for deep feature representation.
  • Utilized FAISS vector indexing to perform fast nearest-neighbor similarity searches across large image datasets.
  • Constructed an interactive Flask web service supporting real-time image upload, feature vector extraction, and candidate retrieval.
Python PyTorch ResNet50 FAISS Flask HTML/CSS/JS
GitHub
AI Complaint Triage & Routing System

AI Complaint Triage & Routing System

  • Developed an automated complaint triage system using TF-IDF vectorization and Logistic Regression to predict target departments and urgency levels.
  • Formulated a hybrid credibility scoring algorithm with dynamic follow-up question generation to validate incoming claims.
  • Engineered a centralized admin dashboard with persistent database storage for auditing, ticket assignment, and oversight.
Python TF-IDF Scikit-learn Flask Database HTML/CSS/JS
GitHub
WasteSnap - Smart Waste Classification

WasteSnap – Smart Waste Classification

  • Trained and fine-tuned a MobileNetV2 convolutional neural network in TensorFlow to classify 9 distinct categories of waste.
  • Delivered real-time inference providing probabilistic confidence scores alongside contextual recycling and disposal recommendations.
  • Packaged and deployed the full-stack web application to production on Render using Flask.
Python TensorFlow Flask MobileNetV2 HTML/CSS/JS
Live Demo GitHub
Mood2Mail Project

Mood2Mail – Email Tone Analyzer

  • Built a natural language classification engine using TF-IDF n-gram vectorization and Naive Bayes modeling to detect email communication tones.
  • Implemented a RESTful Flask API providing instantaneous tone scoring (Formal, Friendly, Aggressive) to the client application.
  • Deployed live on Render with an interactive preview interface delivering real-time sentiment feedback as users type.
Python TF-IDF Flask JavaScript
Live Demo GitHub
Dyslexia AI Assistant

Dyslexia AI Assistant

  • Deployed an assistive text simplification pipeline powered by a fine-tuned T5 Transformer sequence-to-sequence model.
  • Automated the syntactic restructuring of dense, complex passages into simplified sentences while preserving factual nuance.
  • Built an accessible Flask web interface designed with dyslexia-friendly typographic formatting and contrast controls.
Python Flask T5 Transformer HTML/CSS JavaScript
GitHub
DigiDetect Project

DigiDetect – Handwritten Digit Recognition

  • Architected and trained a Convolutional Neural Network (CNN) in TensorFlow/Keras for robust handwritten digit classification.
  • Supported dual input streams with real-time HTML5 Canvas drawing and image file uploads with server-side grayscale normalization.
  • Engineered a Flask backend serving low-latency inference along with full softmax probability distributions.
Python TensorFlow Flask HTML/CSS/JS CNN
GitHub
Real Time ASL Hand Signs Predictor

Real Time ASL Hand Signs Predictor

  • Trained a deep CNN on American Sign Language datasets utilizing GPU acceleration in Google Colab with data augmentation.
  • Integrated MediaPipe and OpenCV pipelines for real-time hand landmark detection and region-of-interest segmentation.
  • Implemented low-latency webcam stream processing to deliver continuous real-time sign language prediction.
Python TensorFlow/Keras CNN OpenCV MediaPipe
GitHub
BiasX-Ray Project

BiasX-Ray – Real-Time Bias Detector

  • Engineered a real-time NLP text analysis engine using spaCy and NLTK to scan communications for gender-coded bias and microaggressions.
  • Implemented a rule-augmented recommendation system proposing inclusive, respectful phrasing alternatives.
  • Deployed full-stack application live on Render featuring asynchronous API communication and visual inline highlights.
Python Flask NLP (spaCy, NLTK) JavaScript
Live Demo GitHub
StyleNet Project

StyleNet – Fashion Image Classifier

  • Designed and trained a deep CNN on the Fashion MNIST benchmark dataset, achieving ~91% test classification accuracy.
  • Implemented server-side image preprocessing pipelines for 28x28 grayscale normalization and tensor reshaping.
  • Deployed as an interactive Flask application enabling drag-and-drop image uploads with immediate categorical predictions.
Python TensorFlow Flask HTML/CSS/JS CNN
GitHub
Music Genre Classification

Music Genre Classification

  • Developed an audio signal processing pipeline leveraging Librosa to extract MFCCs, spectral roll-off, and chroma features.
  • Built a modular Flask REST API handling multiformat audio uploads (MP3, WAV) with configured CORS headers.
  • Optimized audio feature extraction pipelines ensuring compatibility across runtime environments.
Python Librosa Flask API Audio Processing
GitHub
Fake Job Posting Detector

Fake Job Posting Detector

  • Trained a supervised Logistic Regression classifier on labeled job posting datasets to identify fraudulent postings.
  • Executed text preprocessing including tokenization, stop-word filtering, and term frequency feature extraction.
  • Deployed full-stack web application live on Render utilizing a Flask backend and vanilla JavaScript interface.
Python Scikit-learn Flask JavaScript
Live Demo GitHub
Car Mileage Predictor

Car Mileage Predictor

  • Constructed a predictive regression pipeline featuring domain-specific feature engineering and Grid Search parameter tuning.
  • Evaluated multi-attribute automotive specifications to compute calibrated vehicle fuel efficiency metrics.
  • Deployed live on Render with an intuitive form interface and Flask REST backend.
Python Linear Regression Grid Search Flask
Live Demo GitHub
Movie Recommendation System

Movie Recommendation System

  • Architected a content-based recommendation algorithm utilizing cosine similarity matrices across movie titles, genres, and metadata.
  • Built data transformation pipelines in Python to process movie datasets and return ranked high-affinity recommendations.
Python Content-based Filtering Data Analysis
GitHub
+
✦
✦
✦
</>
// docs
PUBLICATIONS

Thoughts, Guides & Books

Sharing practical guides, research breakdowns, and published technical books.

Machine Learning Projects: A Practical Guide for B.Tech CSE Students
Kindle Edition

Machine Learning Projects: A Practical Guide for B.Tech CSE Students

Most machine learning courses teach you theory. This book teaches you how to build. Written by a B.Tech CSE student for B.Tech CSE students, this guide bridges the gap between passing machine learning exams and confidently building projects you can explain and present.

Book Machine Learning Projects B.Tech CSE Kindle
Amazon
+
✦
✦
::
=>
CONNECT

Let's Build Something Great Together

Interested in collaborating or discussing an AI project? Reach out anytime.

Email

bansalmayank1414@gmail.com

Phone

+91 8708574164