Anish
Guntreddi.

Process Engineering Intern

Improved HR process reliability and data consistency across enterprise operations

• Built automation workflows and internal tools for SAP SuccessFactors processes, translating HR technology requirements into reliable scripts, configurations, and software-enabled process improvements
• Collaborated with consultants and stakeholders to identify workflow bottlenecks, gather requirements, troubleshoot system issues, and deliver solutions aligned with enterprise client needs
• Developed validation checks, documentation, and testing workflows to improve process reliability, data consistency, and maintainability across HR operations

AI Engineering Intern

~25% improvement in end-to-end task success, production-grade enterprise AI deployment

• Designed and implemented modular AI agent architectures using LangChain and LangGraph, integrating vector database–backed retrieval and contextual memory to support retrieval-augmented analytics across 5,000+ internal documents spanning multiple enterprise teams
• Developed structured agent evaluation workflows measuring task completion accuracy, response latency, and consistency across retries, benchmarking agent outputs against manual workflows and improving end-to-end task success by ~25%
• Built and deployed full-stack AI systems with FastAPI backends and React frontends to enable real-time agent interaction and secure enterprise deployment, supporting cross-team adoption and production-grade usage within Walmart Global Tech

Software Engineering Intern

60% reduction in manual processing, 20% improved trading efficiency

• Built scalable Python-based automation pipelines integrating Excel, VBA, and SAP APIs to process high-volume enterprise sales orders, reducing manual processing time by ~60% as measured by order-to-completion latency
• Engineered robust validation, logging, and error-handling systems across distributed workflows, reducing data entry errors and operational failures by ~50%, tracked via reconciliation audits
• Designed and deployed automated financial execution systems integrating external trading APIs, improving execution efficiency by ~20% measured through trade execution time and consistency

Machine Learning Researcher

Advancing retinal pathology detection through ML-driven ophthalmic image classification

• Conducted machine learning research on ophthalmic image classification, training and evaluating CNN-based models to detect and classify retinal pathologies from clinical imaging data
• Preprocessed and curated large-scale ophthalmology image datasets, applying augmentation and normalization techniques to improve model robustness and generalization
• Analyzed model performance using clinically relevant metrics (accuracy, sensitivity, specificity, AUC) and collaborated with medical researchers to interpret results in a clinical context

Co-Author and Founding Machine Learning Engineer

~60% directional prediction accuracy, improved predictive stability and Sharpe ratio gains

• Architected a regime-aware Transformer-based ML framework combining deep learning with statistical models (GARCH, HMMs, Monte Carlo), improving predictive stability across volatile market regimes
• Built scalable OHLCV data pipelines with resampling, per-asset normalization, and label smoothing, enabling reproducible experimentation across millions of time-series data points
• Implemented standardized backtesting and evaluation infrastructure, achieving ~60% directional prediction accuracy on out-of-sample data vs. ~52% baseline, resulting in measurable Sharpe ratio gains

Software Engineer

~20% reduction in dropped connections, production-grade rehabilitation app serving thousands of therapy sessions

• Engineered a production-grade Android application in Kotlin, integrating Bluetooth hardware communication, real-time sensor ingestion, and UI feedback loops for rehabilitation devices
• Designed data logging and synchronization pipelines capturing thousands of therapy sessions, evaluated using session completeness, sync reliability, and data integrity metrics
• Implemented mobile architectures supporting offline operation, fault recovery, and secure device pairing, improving session reliability and reducing dropped connections by ~20%

Machine Learning Engineer

~30% improvement in movement classification accuracy, sub-100ms inference latency for real-time rehabilitation

• Designed and deployed an end-to-end machine learning pipeline processing real-time EMG signals, including denoising, segmentation, feature extraction, and inference, enabling adaptive therapeutic assistance with sub-100ms latency
• Designed and trained proprietary ML models to infer patient intent and muscle activation strength from noisy biosignals, improving movement classification accuracy by ~30% over rule-based baselines
• Integrated ML inference outputs with real-time adaptive control logic to dynamically regulate resistance and assistance, supporting personalized rehabilitation across multiple hardware configurations and clinical conditions