Close

Vansh Sachdeva

Software Engineer (AI/ML)

Download Resume

About Me

Results-driven Software Engineer (AI/ML) with 3 years of experience, possessing deep expertise in supervised, unsupervised, and semi-supervised learning algorithms for numerical and structured data. Highly adaptable technical professional with a proven track record of seamlessly executing complex computer vision and Edge-AI video analytics projects. Experienced in architecting scalable backend systems, machine learning analytics, and leading end-to-end project lifecycles from algorithmic design to production deployment.

Experience

Mantra Softech

Software Engineer

  • Key Focus: Production-level deployment, system architecture, and optimization.
  • Advanced ANPR Development: Engineered a multi-layered Automatic Number Plate Recognition system, integrating biometric vehicle verification and facial recognition for high-security environments.
  • Edge & Offline Intelligence: Developed robust local offline synchronization capabilities to ensure 24/7 system availability without cloud dependency.
  • Vision Model Optimization: Refined and deployed YOLOv8 models for real-time safety gear detection, focusing on reducing latency for industrial CCTV applications.
  • NVR Web Application Architecture: Architected and designed a high-performance backend for a Network Video Recorder (NVR) platform utilizing FastAPI (Python) to manage high-throughput video streams efficiently, fully integrated with a responsive, custom-built HTML, CSS, and JavaScript interface.
  • System Integration: Collaborated on building desktop support interfaces and cross-platform biometric solutions. Deep Learning, Natural Language Processing (NLP) and +2 skills

Mantra Softech

Software Developer

  • Key Focus: Production-level deployment, system architecture, and optimization.
  • Advanced ANPR Development: Engineered a multi-layered Automatic Number Plate Recognition system, integrating biometric vehicle verification and facial recognition for high-security environments.
  • Edge & Offline Intelligence: Developed robust local offline synchronization capabilities to ensure 24/7 system availability without cloud dependency.
  • Vision Model Optimization: Refined and deployed YOLOv8 models for real-time safety gear detection, focusing on reducing latency for industrial CCTV applications.
  • System Integration: Collaborated on building desktop support interfaces and cross-platform biometric solutions.

Mantra Softech

Software Engineer Intern

  • Key Focus: Research, model training, and foundation building.
  • Safety Model Prototyping: Conducted initial training and testing of YOLOv8 models for helmet and safety vest detection using custom-curated datasets.
  • Feature Development: Assisted in developing core components for the vehicle verification module and contributed to the codebase for biometric data processing.
  • Data Pipeline Engineering: Built and managed data annotation workflows for training deep learning models, ensuring high-quality ground truth data for computer vision tasks.
  • Technical Research: Explored hardware-accelerated video decoding techniques to improve the efficiency of RTSP stream processing on edge devices.

Independent Developer / AI Consultant

Freelance ML & MERN Stack Developer

  • Formulated solutions and completed a comprehensive MSME use case deck for AIKosh, focusing on practical AI deployments for small and medium enterprises.
  • Architected full-stack e-commerce platforms tailored for European clients using MongoDB, Express.js, React.js, and Node.js.
  • Engineered and integrated ML-driven analytics pipelines to extract actionable business insights from user data and product catalogs.

Education

Center For Development Of Advanced Computing

August 2024 - February 2025

Post Graduate Diploma: Big Data Analytics

Jaypee Institute Of Information Technology

July 2019 - June 2023

Bachelor of Technology - Computer Science And Engineering

Projects

Electrical Energy Consumption Forecasting

Built a time series forecasting model to predict electricity usage with 92%+ accuracy using TBATS, ARIMA, and SARIMA. Processed 150,000+ records and deployed a Streamlit-based web app.

Technologies: Python, Pandas, Statsmodels, Streamlit.

View Project

Algorithmic Trading System

Defined and refined Python-based trading algorithms for NSE/BSE stocks. Implemented technical indicators (SMA50, SMA200, RSI) to execute a two-month swing trading strategy, prioritizing sustained stock appreciation over short-term gains.

Technologies: Python, Pandas.

View Project

Machine Learning Powered Healthcare Management System

Makes the overall work of the healthcare institute easier and provides better results in mental and women health-related fulfilled problems. Capable of serving 1000+ users simultaneously

Tech: JavaScript, ReactJs, TensorFlow, Sklearn, Kaggle, HTML, CSS.

View Project

Mental Health Prediction Using Machine Learning

Developed a project focusing on the crucial aspect of reducing the risk of serious mental disease, which is mental health prediction, and how it can be used as a theoretical foundation to develop plans for behavioural therapies for healthcare workers in the department of public health. Using a database with 10,000 records
Tech: Python, Keras, Intel Extension for Scikit Learn, Scikit Learn, Kaggle
Performance metrics for comparison used are Accuracy, recall And Precision of results

View Project
View More Projects

Publications

Statistical Analysis of Machine Learning Algorithms for Credit Card Fraud Detection

Published at RAMSA Conference 2022

Evaluated models on structured data and identified Random Forest as the optimal algorithm over SVMs.

Skills

Certifications

Get in Touch