SANAT WALIA AI/Ml Engineer
8288961100 codersanat896@[Link] India Linkedin Github
Portfolio Website
Profile
AI/ML Engineer with hands-on experience in MLOps, multi-agent systems, and building production-ready ML pipelines. Known for a
creative hustle — finding smart, innovative ways to get things done and deliver effective solutions.
EDUCATION
DAV MODEL SCHOOL Chandigarh, India
X-92.6% (2018)
XII -95.2%(2020)
Jaypee Institute of Information Technology, Integrated [Link]+ [Link]. Noida, India
CGPA- 7.7
PROFESSIONAL EXPERIENCE
BOBBLE ΑΙ, AI Intern 06/2024 – 12/2024
Published research on face morphing using StyleGANEX in IEEE Xplore (PDGC 2024).
• GURUGRAM, India
Optimized WebP image pipeline, cutting server load by ~33% and improving visual stability.
•
Built diffusion- and GAN-based generative pipelines, enhancing dataset quality and model performance.
•
SKILLS
Programming Languages: Python, C++
Technologies & Frameworks: Streamlit, LangChain, LangGraph, FAISS, PyTorch, Scikit-Learn, OpenCV, FastMCP, Pydantic,FastAPI
Data Science & ML: Pandas, NumPy, NLP,Computer Vision,RAG
Data Processing:: Data Cleaning, Exploratory Data Analysis (EDA), Feature Engineering, Data Preprocessing, Vector Databases (FAISS)
Databases: MySQL, MongoDB
Soft Skills: Creative , Teamwork,Leadership ,Communication
PROJECTS
Revenue Optimization for In-App Purchases, Data Science and Web Application
Developed a full-stack Streamlit app to predict and optimize in-app purchase pricing from user input .
•
Evaluated regression models (Linear, Random Forest, XGBoost) and deployed Random Forest with revenue optimization logic.
•
Delivered an end-to-end pipeline: data cleaning, feature engineering, EDA, model training, evaluation, and deployment with visual
•
insights.
Vidhijna, Legal Research Assistant for Commercial Cases
Built a modular multi-agent system capable of interpreting, retrieving, and analyzing complex commercial legal texts using both
•
vector databases and live web data.
Supports hybrid search: vector retrieval from laws/cases (FAISS) + parallel real-time web research (DuckDuckGo, Tavily, Perplexity).
•
Outputs structured JSON suitable for legal research workflows, supporting custom legal entity extraction and adaptive LLM settings.
•
Designed for extensibility in legal tech applications; demoed for AI-driven jurisprudence analysis.
•
Hotel Reservation Cancellation Prediction - MLOps Pipeline, Machine Learning, Cloud Deployment, CI/CD
Built an end-to-end MLOps pipeline with Python, Scikit-Learn, and LightGBM to predict hotel booking cancellations.
•
Evaluated multiple models and optimized hyperparameters using RandomizedSearchCV, improving robustness and accuracy.
•
Containerized and deployed a modular Flask app on Google Cloud Run with real-time prediction UI and logging.
•
Automated CI/CD with Jenkins and GitHub Actions, reducing deployment time by ~40%.
•
Implemented DVC for data versioning and MLflow for experiment tracking, ensuring reproducibility and transparency.
•
Enhanced model performance through EDA, preprocessing, and feature engineering.
•
AWARDS
Smart India Hackathon 2024 | SIH1670 | WINNERS, Government Of India 12/11/2024
Implemented browser compatibility features to support secure ML model usage, improving accessibility
•
and reliability of the solution.
Collaborated with a cross-functional team to design and integrate technical components, ensuring
•
smooth end-to-end delivery.
Contributed model optimization improvements that enhanced performance and played a role in
•
achieving first place at a national competition.
LEADERSHIP & VOLUNTEERING
AI/ML Hub
Senior Advisor (2024–2025): Mentored AI/ML projects, providing technical guidance and fostering peer learning.
•
Technical Coordinator (2023–2024): Led AI/ML workshops for 60+ students and supported project development through
•
troubleshooting and collaboration.