Foodly — NLP food ordering & tracking assistant
Complete order cycle — browse, place, modify, track — handled in natural language chat instead of a menu UI.
Python · FastAPI · Dialogflow ES · MySQL — source available
I turn unfamiliar problems into working applications — learning what each problem requires, building end-to-end, and improving through iteration.
Curious enough to learn. Driven enough to build.
# what I actually shipSKILLS = { "ml": ["tensorflow", "pandas"], "cv": ["cnn", "opencv"], "nlp": ["dialogflow"], "web": ["fastapi", "django"],}def solve(problem): while not works(problem): problem = learn(problem) return ship(problem)# next: RAG on FAISS + ChromaDB
Complete order cycle — browse, place, modify, track — handled in natural language chat instead of a menu UI.
Python · FastAPI · Dialogflow ES · MySQL — source available
CNN classifies plant disease from a leaf image in seconds, with three disease-specific prevention tips per prediction.
Python · TensorFlow · OpenCV · Django — source available
Tree-based regressor predicts buyer spend from demographics; EDA isolates the segments driving the season.
Python · Pandas · Seaborn · Scikit-learn — source available
AI/ML Developer — curious enough to learn, driven enough to build.
I work across Machine Learning, Computer Vision, NLP and Python development — turning ideas and unfamiliar problems into practical, working applications. I learn what the problem demands, build what it takes, and keep improving until it works.
End-to-end AI systems across analytics, machine learning, computer vision and NLP — from raw data to a running web application. Models ship as interfaces: Django, FastAPI and Flask front-ends over trained TensorFlow and scikit-learn models, with the full workflow hands-on — data cleaning, EDA, feature engineering, training, evaluation and deployment.
Currently pursuing MCA (AI & Machine Learning) at Amity University, 2025 to present, alongside project work.
AI agents, LangChain & OpenAI stack, MCP, OCI platform.
AICTE-approved programme · 8 weeks · modelling end to end.
CGPA 8.15.
Pitched an AI automation solution using facial recognition.
19.5 hours · 88% attendance.
Loops, functions, pointers, arrays · organised by TMU.
Aadharshila The School, Chandpur.
Address correspondence for internships and project work. Messages arrive directly; nothing is stored on this site.
Open to internships, jobs, and project work.