sourabh singh

Designed and implemented 20 complex CLI evaluation environments to stress-test autonomous LLM agents across code generation, environment exploration, and multi-step terminal debugging. Constructed rigorous multi-hop benchmark scenarios exposing model failures in state management, recursive reasoning, and automated error-recovery under constrained resources.

Curated high-accuracy speech datasets for Automatic Speech Recognition (ASR), performing acoustic segmentation, dialect validation, and speaker-turn labeling for multi-modal model training. Executed rigorous S2/S3 LLM alignment evaluations, systematically flagging hallucination vectors, syntax errors, and safety guideline violations to optimize frontier model response quality.

Architected high-throughput AI trading execution agents leveraging Python and TensorFlow to automate real-time multifactor predictive modeling and quantitative strategy execution. Engineered asynchronous event-driven backtesting engines utilizing Redis caching and multithreaded queues, cutting backtest execution latency by 45%.

Developed AI agents for workflow automation, data processing, and business operations. Built n8n-based AI automation pipelines integrating APIs, databases, and third-party services.

Annotated and evaluated LLM outputs to improve model accuracy and alignment across diverse domains. Performed quality assessment, data labeling, and prompt-response validation.

Learned techniques for effective project management including budgeting, scheduling, and resource allocation. Adapted quickly to changing circumstances ensuring that all deliverables were met according to schedule.

Developed machine-learning models for mobile phone price prediction using historical datasets. Applied Logistic Regression, Decision Trees, Random Forests, and SVM for predictive modeling.

Grade: 7.81 CGPA

Pursued a B.Tech in Artificial Intelligence and Data Science, achieving a CGPA of 7.81.

Completed the Google Data Analytics Certification.

Completed the Machine Learning Essentials for Business and Technical Decision Makers certification.