Sanskruti Raut

Hi, I'm Sanskruti Raut

AI/ML Engineer | NLP Researcher | Software Developer

View My Work Get In Touch

About Me

Biography

I'm a passionate AI/ML engineer with expertise in machine learning, natural language processing, and software development. I enjoy building intelligent systems that solve real-world problems and create meaningful impact.

With experience across research and industry applications, I've worked on projects ranging from deep learning models to full-stack applications. I'm constantly learning and exploring cutting-edge technologies in AI and ML.

Download Resume

Skills & Technologies

Python TensorFlow PyTorch Scikit-learn NLP Deep Learning Machine Learning JavaScript React Node.js SQL Git Docker AWS

Education

Master of Science
Electrical Engineering (Machine Learning and Data Science)
University of Southern California
Los Angeles, USA
Aug 2023 - May 2025
Relevant Courses:
Linear Algebra, Probability, Deep Learning, Machine Learning, Natural Language Processing, Cloud Computing
Bachelor of Technology
Electronics and Communication Engineering
Maharashtra Institute of Technology
Pune, India
Aug 2017 - July 2021
Relevant Courses:
Artificial Neural Networks, Data Structures and Algorithms, Computer Networks, Cloud Computing, Internet of Things, Microprocessors and Microcontrollers

Projects

SRGAN Super Resolution
Computer Vision

Computer Vision Super Resolution using SRGAN

Achieved 3x enhancement on DIV2K images with SRGAN (Adversarial Loss: 0.0010, VGG Loss: 0.0014, Pixel Loss: 0.0625)

GitHub →
Transfer Learning Classification
Machine Learning

Transfer Learning for Image Classification

Built a multi-class classifier (ResNet50/101, EfficientNetB0, VGG16), achieving 91.21% F1 and 99.19% AUC with EfficientNetB0.

GitHub →
Laptop Price Prediction
Machine Learning

Laptop Price Prediction using Machine Learning Algorithms

Compared Linear Regression, Random Forest, Support Vector Regression, K-nearest neighbors, and Neural Networks, scoring R² = 0.844 with Random Forest.

GitHub →
Code Detective
NLP

Code Detective

Fine-tuned GraphCodeBERT on NVIDIA A40 for detecting semantically similar Python snippets (F1: 0.96)

GitHub →
Vital Watch
Software Development

Vital Watch

Developed a real-time ICU Patient Monitoring System with custom alerts, displaying vital signs through an interactive dashboard using Node.js/Express backend, React frontend, and PostgreSQL database, deployed on AWS EC2.

GitHub →

Publications

🎓 Google Scholar Profile
Real Estate Based Recommender System Using ML
Raut, S., et al.
GIS Science Journal, 2022
View Paper →
Vehicle Cluster Development
Raut, S., Naware, S., Tank, V.
International Conference on Trends in Computational and Cognitive Sciences (ICTCS), 2021
View Paper →
IOT Based Smart Irrigation System using Cisco Packet Tracer
Raut, S., Motade, S.
International Journal of Computer Sciences and Engineering (IJCSE), 2021
View Paper →

Experience

Aug 2025 - Present
Part-Time

Software Engineer

Easley Dunn Productions Inc
Los Angeles, USA
  • Led the Monster Gridiron Facilities Team by organizing weekly syncs, defining technical milestones, and coordinating backend–frontend integration efforts.
  • Developed and integrated a C#/.NET backend API with a Unity frontend to power facility management features for an American football simulation game.
  • Designed upgrade flows for multiple facilities and implemented dynamic data handling using SQLite and JSON-driven configurations
  • Achieved real-time synchronization between backend database and Unity frontend with API latency under 50ms (local)
Jan 2025 - Present
Research

Natural Language Processing Researcher

Keck School of Medicine, USC
Los Angeles, USA
  • Built a medical RAG system to help HIV clinicians select ART regimens by retrieving evidence from 2024 guidelines
  • Parsed and indexed PDFs using LangChain, BAAI BGE-M3 embeddings, and Chroma vector store for semantic search
  • Integrated and benchmarked multiple LLMs (Qwen-2.5-7B, DeepSeek-7B, GPT-4.1, Gemini 2.5 Pro) with and without RAG to evaluate the impact of retrieval on model performance
  • Developed a full evaluation pipeline using F1, BLEU, ROUGE, and entity-level F1 to quantify performance; best configuration achieved 93.9% accuracy
May 2024 - July 2024
Internship

Machine Learning Engineer

ViyaMD
Los Angeles, USA
  • Conducted a Virtual Product Review (VPR) on multiple Data Ingestion platforms (Azure Data Explorer, Adobe Experience Platform Data Ingestion, PDFMiner, LangChain data loaders, etc.) for Medical RAG use-case
  • Compared extracted PDFs from the VPR with ground-truth PDFs by designing a custom Data ingestion-evaluation pipeline
  • Evaluated pipeline accuracy by measuring precision (92.3%), recall (89%), and F1 (90.12%) scores to quantify extraction and conversion reliability
  • Visualized the above performance analysis using CometML and found Adobe Experience Platform Data Ingestion to be the best
Jan 2024 - April 2024
Research

Machine Learning Researcher

Biomedical Imaging Group, USC
Los Angeles, USA
  • Project 1: Adapted DeepBet v1.0 tool with U-Net architecture for macaque MRI dataset achieving high precision (99%) for skull-stripping
  • Project 2: Trained a PyTorch+MONAI-based SwinUNETR pipeline on NVIDIA A100 GPU, reaching an 80% Dice score for macaque brains for tissue segmentation
Sept 2021 - June 2023
Full-Time

Software Engineer (Analyst)

Deloitte USI
Mumbai, India
  • Created and implemented 150+ automation scripts using Keyword and Data-driven frameworks with Java, Selenium, TestNG, Jenkins, and Maven to test AngularJS based web application with 2026 web pages and performing backend validations with SQL
  • Spearheaded the automation of the regression suite for SIT sign-offs and identifying 200+ defects and reducing manual effort by 60%
  • Collaborated cross-functionally by managing project workflows with JIRA and JAMA, adhering to Agile methodologies and SDLC
  • Recognized with a 'Spot Award' for outstanding automation expertise and contributing to 3 major releases
Dec 2020 - June 2021
Internship

Embedded Software Engineer

Devise Electronics
Pune, India
  • Led a team to develop a digital dashboard using MIL, SIL, and HIL methodology with MATLAB Simulink, successfully enabling real-time monitoring of critical vehicle parameters including speed, RPM, and engine diagnostics
  • Delivered a production-ready, low-cost solution that eliminated parallax errors present in traditional analog systems and featured a modular architecture adaptable across different vehicles with minimal modifications
May 2020 - Aug 2020
Internship

Network Engineer

TATA Communications Pvt Limited
Pune, India
  • Worked as a Network Operations Analyst Intern at TATA Communications' BFSI-SOC division which handles banking customers, focusing on enhancing network incident management efficiency
  • Leveraged ServiceNow and SecureCRT to handle over 100 network tickets with an 85% first-call resolution rate and developed streamlined troubleshooting workflows for MPLS/BGP network faults, reducing average resolution time from 25 to 5 minutes
  • Developed a Python-based analysis tool using pandas and matplotlib to analyze patterns across 1000+ historical network incidents, which helped identify and reduce repeat failures by 30%

Get In Touch

I'm always open to discussing new projects, opportunities, or collaborations.

📧
sanskrut@usc.edu
💼
LinkedIn
🐙
GitHub