Hello!
Vaibhavi Deshpande
Software engineering, Data science, Machine learning
"Building reliable systems and intelligent workflows that transform data into actionable insights."
About Me
Building Impactful, User-First, and Intelligent Solutions
Core Principles
- Design systems that reduce friction for users while delivering actionable insights
- Bridge robust software engineering with pragmatic AI and data-driven workflows
- Solve real-world problems using ML, analytics, and automation
- Embrace continuous learning and emerging technologies to drive innovation
Technical Skills
Programming & Scripting: Python (Pandas, NumPy, scikit-learn), C++, JavaScript, SQL
Machine Learning & AI: TensorFlow, PyTorch, OpenCV, LangChain, Matplotlib, Seaborn, spaCy, NLTK, RAG
Database Management: PostgreSQL, MySQL, SQLite, MongoDB
Tools & Technologies: Git, GitHub, Docker, Kubernetes, AWS (S3, RDS), FastAPI, Postman, Power BI, Tableau
Certificates: AZ-900 Microsoft Azure Fundamentals, Google Analytics
Experience
Allyzent
Business and Engineering Solutions Team Inc.
University of Central Florida
PALASH Healthcare
Kaizen Chemtech
AI Engineer
Built scalable data ingestion and preprocessing pipelines
- Developed ML-based confidence scoring models to evaluate prediction reliability
- NLP-based sentiment extraction from financial news using transformer-based embeddings
Allyzent | Jan 2025 - Present
Software Engineer
ERP Data Pipelines, APIs & Predictive Maintenance
- Engineered Python-based ETL with Pandas & SQL to clean and standardize ERP data.
- Built FastAPI services to power real-time Power BI dashboards for maintenance and orders.
- Implemented predictive maintenance models (scikit-learn, XGBoost) for early failure alerts.
- Designed Power BI & Looker dashboards for KPI monitoring
BEST Inc. | June 2025 - Dec - 2025
Education
Master of Science — Computer Science
University of Central Florida | GPA: 3.9
Aug 2022 — May 2024
Bachelor of Technology — Electronics & Telecommunication
Pune University
Aug 2018 — May 2022
Project Showcase
AI-Powered Order Validation
FastAPI backend with an AI agent (Claude) to parse emails and validate orders; deployed with Docker on AWS.
GitHub
Multi-Agent Medical Intelligence
Claude LLM with LangChain/LangGraph and RAG to retrieve PubMed evidence; FastAPI microservices orchestrate agents.
[ Repo / Demo ]
Retail Recommendation System
BERT embeddings with K-Means clustering for personalized product recommendations.
GitHub
Handwritten Character Recognition
CNNs on EMNIST + 10k custom samples; noise reduction, normalization, feature extraction; 98% accuracy.
Credit Card Fraud Detection
Vesta dataset, feature engineering and tuned XGBoost; 96.7% accuracy for secure transactions.
GitHub
Detecting Fraudulent Job Postings
BERT/DistilBERT-based classifier with optimized hyperparameters; 99.049% accuracy.
Beyond the Code
Skydiving is next!
Embracing Challenges
My love for adventure sports like scuba diving, rafting, and bungee jumping isn't just a hobby—it's part of my mindset.
This mindset of embracing new challenges, staying calm under pressure, and pushing boundaries is something I bring to my professional life every day. It keeps me curious, driven, and always ready for what's next.