Thriyogya K

About Thriyogya K

AI Engineer with around 6 years of experience designing and delivering end-to-end Machine Learning and Generative AI solutions in enterprise environments, specializing in translating complex business requirements into scalable, production-ready AI architectures. Strong foundation in advanced statistical modeling, supervised and unsupervised learning, deep learning (CNNs, LSTMs, Transformers), and NLP-driven systems, with hands-on expertise in building LLM-powered applications such as embeddings, semantic search, and Retrieval-Augmented Generation (RAG) pipelines. Proficient in Python, SQL, distributed data processing with Spark, and cloud AI ecosystems (AWS, Azure, GCP), with extensive experience in data preprocessing, feature engineering, model optimization, and hyperparameter tuning to enhance performance and generalization. Well-versed in MLOps practices including model versioning, CI/CD automation, Docker containerization, and performance monitoring, along with a strong commitment to model explainability, bias detection, and responsible AI principles. Proven ability to collaborate with cross-functional stakeholders to deliver scalable, reliable, cost-optimized AI systems that drive measurable business impact.

Experience

1. Ai Ml Engineer

   IBM

   Jul 2024 — Present · US

   Designed and built end-to-end AI solutions from raw data ingestion to model deployment solving real business problems using Machine Learning and Deep Learning techniques.•Developed scalable ML pipelines using Python, leveraging libraries like NumPy, Pandas, Scikit-learn, and deep learning frameworks such as TensorFlow and Py Torch.•Worked extensively on supervised and unsupervised learning models including regression, classification, clustering, ensemble methods (Random Forest, XG Boost), and neural networks (CNNs, RNNs, LSTMs).•Designed and fine-tuned NLP models using Transformers, Hugging Face, and implemented LLM-based solutions including prompt engineering, embeddings, RAG pipelines, and vector databases.•Built and optimized data pipelines using SQL, Spark, and Airflow, ensuring efficient data preprocessing, feature engineering, and model reproducibility.•Performed advanced feature engineering, handling missing data, outlier treatment, encoding techniques, and dimensionality reduction (PCA, t-SNE) to improve model performance.•Implemented ML Ops best practices: model versioning (ML flow), containerization (Docker), CI/CD pipelines, and deployment on cloud platforms like AWS, Azure, or GCP.•Deployed production-grade APIs using Fast API or Flask, enabling real-time inference with low latency and high availability.•Optimized models through hyperparameter tuning (Grid Search, Random Search, Bayesian optimization) and improved performance using cross-validation strategies.•Monitored production models for drift, bias, and performance degradation, implementing retraining strategies and automated alerts.•Worked closely with product managers, data engineers, and business stakeholders to translate business requirements into measurable AI-driven outcomes.•Ensured explainability and fairness using SHAP, LIME, and bias detection frameworks to maintain responsible AI standards.

Education

• University of New Haven

  Master's Degree, Data Science

• Vignan's Foundation for Science, Technology & Research

  Bachelor's Degree, Electrical, Electronics and Communications Engineering