Nitin Lodha

* Building a GUI to process 125 Hz physiologic waveform data for CA and CVR analysis in longitudinal LVAD studies. * Implementing mX autoregulation and CVR metric computation using windowed averaging, sliding correlations, and CO₂ response modeling. * Building automated data cleaning, epoch selection, and structured export pipelines for reproducible analysis.

• Applied natural language processing (NLP) and machine learning techniques to analyze social media data on grief, violence, and healing. • Collaborated with interdisciplinary teams to develop frameworks identifying emotional distress and community resilience signals. • Supported the lab’s mission of advancing ethical AI and amplifying marginalized youth voices online.

• Designed and deployed solutions with the revenue management team to address 5G telecom challenges, improving billing accuracy and network reliability for high-volume clients. • Developed and implemented a custom code generation system that automated test code creation for new use cases, reducing manual coding by 70% and accelerating validation cycles by 3×. • Built and launched a pre-meeting planner tool that generated contextual knowledge bases from past meetings, enabling faster onboarding for participants and boosting productivity by 35%.

• Solved critical 5G telecom challenges, troubleshooting defects and building features that improved system reliability by 20%. • Developed a customized command recommender system, reducing manual lookup by 60% and troubleshooting time by 40%. • Delivered end-to-end contributions, development, testing, and trouble report resolution, achieving 95% on-time delivery and boosting customer satisfaction.

• Researched 10K+ Indian POI utterances, expanding the dataset coverage by 35% for better representation. • Built an ML model for Indian POI Auto-NU creation, improving utterance diversity and accuracy by 40%. • Benchmarked against NLP models (BERT, T5), achieving ~25% higher performance and generating more contextually diverse utterances.

• Engineered a centralized Splunk dashboard that streamlined analysis of 100K+ ERROR/WARN logs, enabling teams to detect and resolve issues 35% faster. • Designed and deployed 10+ dynamic visualizations (graphs, filters, trend analysis) that improved cross-team monitoring, decision-making, and reduced troubleshooting overhead.

• Boosted user engagement by 25% and expanded active userbase through improved website stickiness and usability. • Designed user personas and implemented optimized UI/UX flows, reducing bounce rate by 15%. • Built the landing page and resolved 20+ backend issues, enhancing site stability and page load time by 30%.

Coursework: CIS 5200 Machine Learning CIS 5480 Operating Systems Design and Implementation CIS 5810 Computer Vision & Computational Photography CIS 5020 Analysis of Algorithms CIS 5230 Ethical Algorithm Design STAT 5850 Deep Learning with Applications