Ummusalma P T

About

AI engineer with hands-on experience building end-to-end AI applications across machine learning, computer vision, RAG, agentic AI, and LLM-powered systems. Skilled in developing intelligent applications and production-ready AI solutions using Python, FastAPI, Django, PostgreSQL, Docker, and MLOps practices.

Experience

• Data Science Intern
  SMEC Automation Pvt. Ltd.
  Dec 2025 – Feb 2026

Education

• B.Tech
  APJ Abdul Kalam Technological University
  Computer Science · 2021 – 2025

Skills

• SQL
• FastAPI
• Git
• Python
• Power BI
• Prompt Engineering
• MongoDB
• Django
• Feature Engineering
• Matplotlib
• NumPy
• Pandas
• scikit-learn
• OpenCV
• PostgreSQL
• PyTorch
• TensorFlow
• Seaborn
• Docker
• Django Rest Framework
• Statistical Analysis
• REST APIs
• OCR
• Agentic AI
• RAG
• Anomaly detection
• Pytest
• Object Detection
• CNN
• LangGraph
• GitHub Actions
• Embeddings
• MLflow
• Hugging Face Transformers
• Eda
• Sentence Transformers
• Faiss
• Chromadb
• Image Classification
• Groq
• Yolov8
• object tracking
• Llama 3.1
• Crossencoder
• Railway

Projects

• Real-Time Road Accident Detection & Alert System
  YOLOv8, OpenCV, FastAPI, Streamlit

  Engineered a real-time road accident detection and emergency response system using YOLOv8, OpenCV, object tracking, and severity scoring to automatically identify accidents, generate alerts, and provide live monitoring through FastAPI services and Streamlit dashboards.

• AI-Powered Semantic Job Matching System
  FastAPI, Sentence-BERT, FAISS, CrossEncoder

  Designed a semantic job matching platform using Sentence-BERT embeddings, FAISS vector search, CrossEncoder reranking, and LLM-powered career insights to deliver personalized job recommendations, skill-gap analysis, and retrieval quality evaluation using Hit@K, MRR, and Recall@K metrics.

• Privacy Policy Risk Analyzer
  FastAPI, Groq LLaMA 3.1, PostgreSQL, MLflow, Docker, PyTest

  Built an end-to-end AI-powered privacy policy analysis platform using DeBERTa-v3 and Groq LLaMA 3.1 to identify risky clauses, generate explainable risk assessments and recommendations, and enable production deployment through FastAPI, PostgreSQL, MLflow monitoring, Docker, PyTest, and GitHub Actions CI/CD.

• AI Personal Finance Copilot
  Django, PostgreSQL, Groq LLM, RAG, ChromaDB

  Developed and deployed an AI-powered personal finance assistant using Django, Groq LLaMA 3.1, ChromaDB RAG, and PostgreSQL, integrating agentic financial Q&A, OCR-based receipt parsing, anomaly detection, expense forecasting, and automated PDF reporting.

Courses & certifications

• Introduction to Agent Skills · Anthropic
• Power BI · SMEClabs
• Data Analytics · SMEClabs
• Professional Certification in AI & ML with Data Science · SMEClabs

🗣️ Languages

• English · Fluent
• Malayalam · Native
• Tamil · Basic