Machine Learning Ai Engineer

SIDDHARTH RANJAN

********@***.*** 415-***-**** https://www.linkedin.com/in/mr-siddharth-ranjan/

PROFESSIONAL EXPERIENCE

BLUU KAZI Oxon Hill, Maryland, AI Engineer Aug 2024 - Present

● Led AI engineering to build a semantic matching engine and PDF parser using Gemini text embedding 004, Gemini 1.5 Flash, and NLP to improve accuracy and reduce errors.

● Implemented similarity scoring with a dynamic, field weighted system in Python and SQL, letting recruiters prioritize skills, experience, and certifications for tailored fit scores.

● Built an AWS pipeline with S3, Lambda, SQS, DynamoDB, PostgreSQL, and FastAPI, collaborating with the frontend team using Rust for seamless integration and low latency performance.

● Containerized inference services using Docker and deployed via GitLab CI/CD on AWS, integrating a vector database for embeddings and ensuring scalable, fault tolerant operation.

● Created a JSON fallback for malformed and incomplete PDFs, ensuring structured outputs; integrated Airflow and CloudWatch for automated monitoring and alerts.

● Monitored and optimized pipelines using performance metrics, error tracking, and logging to drive continuous improvements in accuracy, speed, and reliability.

● Optimized parsing and pipeline execution to cut processing time by 30%, enabling faster and more reliable hiring evaluations.

PROJECTS

Wealth Genie: Portfolio Predictor Chatbot

● Conducted predictive modeling with LSTM and sentiment analysis using FinBERT, automating portfolio adjustment decisions for improved investment outcomes.

● Integrated NLP and RAG with LLaMA and FAISS, enabling context aware and intelligent financial question answering for investment related queries.

● Visualized performance trends by applying advanced data analysis and interactive visualization tools by using D3.js, creating clear and insightful dashboards.

PhishScan

● Evaluated nine machine learning models (XGBoost, SVM, Random Forest) on a phishing dataset using Python and scikit learn. Achieved 97% accuracy through tuning and cross validation, boosting detection reliability.

● Leveraged Boruta feature selection, BERT embeddings, and ensemble strategies for enhanced performance, minimizing false positives.

● Applied data mining methods and deep learning principles to optimize feature extraction and improve classification robustness.

New York Airbnb Price Prediction System

● Preprocessed the NYC Airbnb dataset using KNN imputation and categorical encoding, improving data quality by 15% for robust predictive modeling.

● Applied hypothesis testing using Mann Whitney U, Kruskal Wallis, and t tests to identify statistically significant pricing differences across boroughs and room types.

● Built a Django web application integrating the predictive model, storing data in MongoDB, and visualizing pricing trends using Tableau.

TECHNICAL SKILLS

Languages: Python,SQL, R, HTML, CSS, JavaScript AI & ML: LLM, Generative AI, Neural Networks, Transformers, CNN, RNN, ANN, SARIMA, ResNet18, Swin Transformer, UNet, Faster RCNN, PyTorch, TensorFlow, Keras, OpenCV, NumPy, Pandas, Seaborn,scikit learn,cross validation,Boruta CI/CD & Containerization: Git, GitLab, GitHub, Docker Databases: MongoDB, NoSQL, MySQL, ZOHO,Oracle Cloud & Big Data: Amazon Web Services, Google Cloud Platform, Groq, Apache Spark, Airflow, Snowflake, Redshift,Databricks Analytical Tools: Power BI, MS Excel, Tableau, Looker EDUCATION

Arizona State University (3.45/4)Tempe, Arizona

Master of Science in Data Science, Analytics and Engineering CERTIFICATIONS

Data Visualization with Tableau, Google Data Analytics Professional Certificate,Data Engineering, Big Data, and Machine Learning on GCP,Get started quickly with Jira,GitLab with Git Essentials

Contact this candidate