• Advised on AI modeling approaches and evaluated project designs for feasibility, scope, and business impact across multiple AI domains, including GenAI, anomaly detection, and forecasting.
• Prioritized AI use cases across predictive maintenance and battery lifespan monitoring, reducing battery failure risk by 19% and accelerating testing cycles by 6%, enabling faster launches for next-generation vehicle models.
• Designed agentic AI system ingesting XML, HTML, and datalake tables, leveraging MCPs to autonomously validate battery data and generate actionable insights, reducing manual data validation by 40%.
• Mentored cross-functional team of 5 engineers on AI best practices, establishing standardized approaches for model benchmarking and deployment readiness across the organization.
• Acted as primary strategic interface between technical delivery and executive leadership, directing agile delivery across technical team through sprint planning and backlog prioritization.
• Established evaluation framework with KPIs and complexity scoring, translating findings for executives and non-technical stakeholders to enable C-level roadmap decisions and cross-functional alignment.

• Led end-to-end EV data pipeline modernization, migrating legacy ETL architecture from proprietary tools to open-source stack, achieving 40% improvement in processing speed and 50%+ cost reduction while maintainin zero downtime.
• Architected a scalable AI platform with LLM-powered RAG chatbot for natural language querying of 500K battery test records, enabling 40+ engineers to self-serve instead of relying on 2 data specialists.
• Conducted presales activities, including technical requirements gathering and solution design, evaluating multipl technologies, through POCs to recommend optimal semantic search architecture aligned with client needs.
• Integrated CI/CD workflows and deployment automation for data infrastructure, ML models, and analytic components, reducing manual operational overhead by 35%.
• Mentored engineering team of 3 on modern data engineering practices, LLM evaluation methodologies, and data governance, translating complex AI capabilities for non-technical stakeholders to secure buy-in for continued investments.

• Fine-tuned a domain-specific LLM to improve machine translation accuracy for automotive industry jargon, enhancing multilingual communication across internal systems.
• Managed end-to-end delivery of OCR automation solution, directing the digitization and structured extraction of data from scanned business documents to optimize operational and compliance workflows.
• Delivered two production-ready AI services with 30% performance gains over baseline models, deployed as Saas products for internal and external client use.

? Designed an AI system for social media threat detection, combining multilingual NLP and computer vision to improve early identification of criminal patterns.

? Led cross-functional PoC reviews with academic and industry stakeholders, aligning technical direction with real-world use cases and regulatory requirements.

? Developed automated pipelines for multilingual text and image preprocessing, reducing manual labeling time by 40% and enabling scalable training workflows.

? Engineered and trained custom neural networks for cross-modal classification, achieving 20% improvement in prediction accuracy and 20% higher clustering performance using unsupervised embeddings.

• Architected a real-time data processing solution for driver assistance system testing, enabling remote access and touch-command simulation workflows to streamline validation processes.
• Led agile collaboration between developers, test engineers, and client stakeholders to manage sprint planning, requirements tracking, and ensure full alignment with business objectives.
• Delivered multiple cloud-native deployments, supporting remote testing and laying the foundation for AI-based analytics, predictive validation, and automated defect detection.

• Designed and implemented a scalable data processing pipeline for bacterial genome research, enabling 100+ doctoral students to manipulate and analyze gene expression data efficiently.
• Integrated ETL workflows with statistical analysis tools to automate preprocessing, feature extraction, and transformation tasks, reducing manual effort by 40% and accelerating time to insight.
• Enabled downstream use of AI models by standardizing datasets and supporting reproducible research workflows across multiple use cases in bioinformatics.