Research Assistant · AI & Machine Learning
I build with AI, data, and the cloud.
Research in AI & ML at Concordia, with a thing for LLMs, computer vision, and turning messy data into something useful. When I'm not in the lab, I'm probably breaking something in the cloud—on purpose.
I'm Bhavithra—I work at the messy middle between research and real-world systems.
After my Master's in IT at Concordia (Edmonton), I stayed on as a Research Assistant in AI & ML. Most of my time goes into LLMs, multimodal models, and a lot of Python. Before that I was in data: building pipelines, dashboards, and making sure the right numbers reached the right people.
I care about problems where “it works in the lab” isn’t enough—getting a model to run reliably in the wild is the fun part. I use cloud, APIs, and a bit of full-stack to close that gap.
When I’m off the keyboard: sci-fi, tech reads, and the occasional sunset.
How I work
The same rhythm whether it is a model, a dashboard, or a class full of students: frame the problem, iterate where it is measurable, then make it easy for the next person to run.
Clarify success metrics, data constraints, and failure modes before touching architecture—so experiments stay honest.
Prototype fast, log everything useful, and compare against baselines. Favor small, reviewable changes over big-bang rewrites.
Docs, reproducible runs, and deployment paths matter. The best work is boring to operate and exciting to extend.
Research, teaching, analytics, and cloud operations.
Concordia University of Edmonton, Edmonton, AB
Concordia University of Edmonton, Edmonton, AB
Besant Technologies, Chennai, India
BinaryLogic, Edmonton, AB
AI/ML, full stack, and data work.
Multimodal deep learning for debris detection and tracking—validated against a strong baseline (thesis benchmark).
Geo-location attendance system for 500+ users. JWT, RBAC, dashboards. 75% faster reporting.
Real-time snow detection with YOLOv8. 91% accuracy, 50+ FPS, Flask deployment.
Tkinter GUI chatbot with sentiment analysis, trivia quiz, and personalized responses.
Regression model to forecast CO2 emissions from historical and environmental data.
Database system for student records, courses, faculty. PHP web interface.
Power BI dashboard for user behavior, content trends, revenue.
Regression model for airline delays from weather and traffic data.
Classification model for exoplanet characteristics from astronomical features.
Client-server apps with sockets. TCP/IP, inter-process communication.
Stack and tools I use regularly.
August 2024 – December 2025
GPA: 3.92/4.0
July 2017 – October 2021
CGPA: 8.3/10
Open to research collabs, interesting problems, or a good tech chat. Say hi.