Languages
Python, Java, C++, TypeScript, SQL
ANUJ WADI
Software Engineer | AI/ML Engineer
Who am I?
I'm a Software Engineer and AI/ML Engineer with an MS in Robotics and Autonomous Systems (AI) from Arizona State University. I like building things that actually get used, not just demoed. I built an AI-assisted Blood Bank Management System that's now running in 3 hospitals, and a multimodal AI agent that pulls from live data sources so it answers from real information instead of guessing. Most of what I do lives at the intersection of agentic AI, applied machine learning, and robotics: writing the backend, training the model, and getting it into someone's hands.
Technical Skills
Core CS
Data Structures & Algorithms, Object-Oriented Programming (OOP), ROS, Motion Planning
Agentic AI
MCP, LLM APIs, Agentic Workflows, Prompt Engineering
Backend & Cloud
REST APIs, Pandas, AWS, GCP, Docker, Kubernetes, Terraform
Machine Learning
Scikit-learn, Random Forest, Gradient Boosting, Regression Models
Languages
Scroll to explore โ
Experience
AI Full Stack Engineer
PrimaThink Technologies Limited
Jan 2024 โ Jun 2024
- Built an AI-assisted Blood Bank Management System with a 5-person team and rolled it out to 3 hospitals โ interviewed 12+ coordinators to get the workflow right, then trained 20+ staff to replace their manual inventory tracking.
- Trained Random Forest and Gradient Boosting models on 10K+ donation records to forecast blood shortages 7 days out, so hospitals could plan donor drives before they actually ran short.
- Set up CI/CD with 85%+ Jest coverage, which brought bug turnaround down from 2 days to under 4 hours.
Web Development Intern
PrimaThink Technologies Limited
Jul 2023 โ Dec 2023
- Took an early-stage startup's first website from nothing to 12K+ users, owning everything from design to deployment.
- Built the 25-page responsive site on JavaScript, React, and PHP, working closely with 3 founders and a designer, and ran biweekly demos to keep the launch backlog on track.
- Got promoted to AI Full Stack Engineer after the launch, moving from front-end work into machine learning.
Machine Learning Intern
Tech Learn. Live
Jul 2021 โ Dec 2021
- Ran 6 ML experiments and pushed up accuracy on each โ including a linear regression model for stock price movement and a symptom-based heart attack risk detector.
- Trained and evaluated models in Python, Scikit-learn, and TensorFlow with a team of 3, testing classification, regression, and neural-network approaches to find what actually worked best for each one.
- Built reusable data pipelines (feature engineering, normalization, cross-validation) that the rest of a 4-intern cohort ended up using too, and reported findings to 2 senior data scientists every week.
Education
Master of Science, Robotics and Autonomous Systems (AI)
Arizona State University
Aug 2024 โ May 2026
- Coursework in machine learning, computer vision, and robotics, focused on building autonomous systems that hold up outside a simulation.
- Hands-on projects in AI-driven control systems, multi-agent coordination, and sensor fusion.
- Working on research into autonomous systems that need to be reliable in practice, not just in theory.
Bachelor of Technology, Major in Artificial Intelligence
G. H. Raisoni Institute of Engineering and Technology
Aug 2020 โ Jun 2024
- Core coursework in machine learning, deep learning, NLP, and data structures.
- Led academic projects on mental health monitoring and stock price prediction.
- Worked as a grader and project lead โ a lot of writing docs and helping other students debug their code.
My Work
Multimodal AI Agent Runtime
An agent that pulls from 5 live knowledge sources via RAG before responding, so answers stay grounded instead of drifting into generic LLM territory. A background job queue cut response wait from 20s to 7s. Built with Python, FastAPI, Claude, RAG, and ElevenLabs.
20s โ 7s Wait Time
FlipFlick
A movie-decision app disguised as a slot machine โ pick a vibe, get matched by a custom TF-IDF recommender scored on similarity, rating, and popularity. A softmax sampler keeps every reroll feeling fresh instead of repeating. Built with Python, FastAPI, scikit-learn, and Next.js.
ML-Powered
View Source โ
Harmony โ Mental Health Monitoring for Astronauts
Led a team of 5 building a mental health monitoring platform for astronauts โ LSTM and BERT models flagging distress signals, encrypted storage, and inference cut from 1.4s to 0.9s. Built with Python, TensorFlow, LSTM, and BERT.
87% Accuracy
Maze Navigation with MyCobot Pro 600
Autonomous maze-navigation system using ROS, OpenCV, and inverse kinematics โ 92% success rate across 50+ trials. A Gazebo simulation cut iteration cycles from 3 hours to 45 minutes.
92% Success Rate
Parksnese - Smart Parking Solution
Built an ML-based parking prediction system with a full IoT stack behind it, and pitched it to 3 investors. Cut parking search time by 40% in testing.
Chalkboard AI
Built a working website for an AI chalkboard in 3 weeks using Lovable AI and OpenAI โ real-time transcription and summarization to help students actually keep up in class.
Stock Price Predictor | Published Research
Built a Q-learning model that processes 50K+ data points a day to generate trading signals โ backtested to a 17% simulated ROI increase over six months.
Published Research
College Admission Predictor
Built a prediction model with Random Forest and Logistic Regression that helped 129+ students find schools where they had better odds of getting in.
Heart Disease Prediction
Built an ML model with Random Forest and SVM that hit 91% accuracy, using an Apache Spark pipeline to handle the large medical datasets.
91% Accuracy
Career snapshot
Resume
Last updated August 2026
Preview unavailable in this browser.
Use the download button to view the resume.
Contact Me
Send a Message
Get In Touch
Have a project, a role, or just a question? Reach out โ I read every message and I'm quick to reply.
Email
Location
Arizona, USA โ Open to Relocation
Work Authorization
F-1 STEM OPT โ authorized 3 years, no sponsorship required to start