Developed deep learning models for financial time-series forecasting using TensorFlow and PyTorch with GPU-accelerated training, improving asset price prediction performance by 22% and supporting algorithmic trading and portfolio analysis.

Implemented NLP pipelines and fine-tuned BERT transformers via Hugging Face to analyze financial news and social media sentiment, and integrated RAG retrieval mechanisms for enriched context, enhancing trading signal precision by 18%.

Built scalable machine learning workflows on GCP using BigQuery and Vertex AI, enabling distributed data processing and deployment of inference services for AI-driven financial analytics.

Designed anomaly detection systems using autoencoders and unsupervised learning techniques to identify suspicious transaction patterns, optimizing ROC-AUC thresholds to reduce false positives by 30%.

Implemented CI/CD pipelines for machine learning models using Docker, Kubernetes, and GitHub Actions, enabling automated model training, testing, versioning, and deployment while reducing release cycles by 40%.

Partnered with quantitative analysts, traders, and compliance teams to ensure models adhered to regulatory standards and business KPIs, improving transparency, reliability, and stakeholder trust through LLM-driven interpretability frameworks.

Developed deep learning models using Convolutional Neural Networks (CNNs) for medical imaging analysis, leveraging GPU-enabled training in PyTorch and TensorFlow to improve diagnostic model accuracy to 94% validation performance.

Engineered multiple behavioral and physiological signal features using embedding similarity techniques and statistical feature extraction, improving model precision by 38%.

Designed real-time gaze tracking algorithms using computer vision techniques in OpenCV combined with linear transformations and normalization methods to ensure consistent predictions across devices.

Built machine learning pipelines on AWS using Lambda, Glue, and S3, enabling automated data ingestion, preprocessing, model retraining, and deployment workflows.

Collaborated with healthcare researchers to convert clinical diagnostic parameters into machine learning features, integrating NLP-based text processing to analyze clinical reports and medical documentation.

Developed Power BI dashboards connected to SQL databases to monitor model performance metrics such as inference latency, ROC-AUC, and prediction drift, enabling data-driven improvements for research teams.

Performed exploratory data analysis (EDA) using Python, Pandas, NumPy, and Matplotlib on multi-source enterprise datasets, identifying data quality issues and feature patterns that improved downstream model performance by 14%.

Created time series forecasting models using ARIMA-based techniques and Keras to predict demand and operational metrics, reducing forecast error (MAPE) by 18% and supporting short-term capacity planning.

Executed supervised and unsupervised learning models, including Decision Trees, Random Forests, K-Means, DBSCAN, and PCA using Scikit-learn, increasing classification accuracy by 16% and enabling customer segmentation for business analytics use cases.

Processed large-scale datasets using PySpark and SQL Server, and delivered analytical insights through Power BI dashboards, accelerating reporting cycles by 35% and improving visibility for project stakeholders.