Ravi

AI Engineer and Data Scientist with 5+ years of experience delivering production-grade machine learning, generative AI, and data science solutions across healthcare and industrial sectors. Led development of advanced analytics like Retrieval Augmented Generation and predictive models to business objectives, with hands-on experience across the full stack from database architecture and data pipelines to cloud deployment. Works cross-functionally with medical affairs, product, and operations stakeholders to translate customer needs into go-to-market AI capabilities.

Strong background in SaaS platform development, CRM data integration, business analytics, and automation tools.

Microsoft Certified Azure Data Scientist

Associate and Oracle Generative AI Professional.

AI Engineer, Generative AI and RAG - Axtria - Remote

(2024-09)

• Built end to production RAG platform to help medical teams quickly search 15,000+ clinical documents, cutting research time from 15 minutes to under 30 seconds and improving retrieval accuracy to 95%@10.
• Designed a hybrid retrieval system combining BM25 lexical search with semantic embeddings and optimized chunking (512 tokens, 25% overlap) to serve accurate, context-aware results aligned to customer needs.
• Developed full-stack FastAPI services with LangGraph multi-agent orchestration, input validation, and observability hooks, creating a standardized AI service deployed across Azure and AWS.
• Improved clinical query accuracy from 78% to 94% using prompt engineering with chain-of-thought reasoning and few-shot techniques, reducing physician query iterations by 60% and enabling business objectives around decision-making efficiency.
• Led document understanding pipelines using PyMuPDF and Azure Document Intelligence feeding AWS Bedrock-powered workflows, applying automation tools to accelerate information extraction at scale.
• Built database ingestion and governance frameworks covering schema validation, metadata freshness, and automated index refreshes, enabling reliable business analytics reporting across the content ecosystem.
• Developed statistical drift detection and predictive model monitoring for monthly datasets, publishing segmented performance reports to support proactive model governance aligned to business KPIs.
• Implemented CI/CD pipelines and containerization using Docker and GitHub Actions, applying infrastructure-as-code principles to reduce troubleshooting time and ensure consistent releases across environments.
• Designed and implemented data governance frameworks on GCP Data Catalog that included schema validation, metadata management, data lineage tracking, and automated index refresh policies for a 15,000+ document clinical content ecosystem, improving data traceability and compliance.
• Mentored two junior engineers on the team, reviewing their RAG implementations and guiding trade-off decisions on latency, accuracy and re-indexing cost, which improved deployment efficiency and reduced model response latency.

Data Scientist, Machine Learning and Computer Vision - Tech Mahindra - Hyderabad, India

(2020-06 - 2023-01)

• Designed end-to-end predictive models for industrial fault detection framed as supervised classification, delivering Random Forest models at 88% precision and 87% recall with metrics tied directly to business objectives.
• Reduced false positives by 12% through adaptive threshold tuning using multi-armed bandit reinforcement learning, improving automated inspection reliability across advanced analytics-driven quality control workflows.
• Built and deployed computer vision systems using YOLOv3 fine-tuned on 50,000+ images, achieving 92% mAP and sub-500ms inference latency for real-time service delivery in quality inspection.
• Designed real-time Kafka and MongoDB database ingestion pipelines supporting low-latency model scoring and downstream business analytics consumption across manufacturing operations.
• Built end-to-end ML pipelines in Python using scikit-learn and Pandas for data validation, preprocessing, feature engineering, model training, and evaluation, which streamlined the workflow and ensured consistent, reproducible model development.
• Deployed containerized ML services on Kubernetes with autoscaling and rolling updates, applying infrastructure-as-code and troubleshooting practices to maintain 99.8% uptime across 50,000+ daily inference requests.

Jr. Data Scientist - Microland - Bengaluru, India

(2019-01 - 2020-06)

• Queried and joined customer data from MS SQL Server databases using SQL, building analysis-ready datasets and performing data quality checks that fed business analytics dashboards and executive reporting.
• Built customer segmentation predictive models using PCA and K-Means clustering, validating outputs through silhouette scoring and informing CRM targeting strategies and go-to-market planning.
• Conducted statistics driven root cause analysis on hardware test data using SciPy and Statsmodels, identifying failure patterns for 15% of recurring issues and delivering recommendations adopted by the engineering team.
• Automated data collection and validation workflows using Python and Bash scripting, reducing hardware validation cycle time by 30% through targeted use of automation tools aligned to business objectives.
• Automated recurring data collection via Python web scraping with BeautifulSoup, converting unstructured web pages into structured datasets for analytics.
• Performed exploratory data analysis using Python and pandas to identify patterns, outliers, and key drivers, applying missing-value treatment, normalization, and feature scaling, which informed feature selection for predictive models and improved forecast accuracy.
• Designed Power BI and Tableau dashboards translating business models into visual reporting used by product marketing and operations leaders across business units.

Master of Science - Computer and Information Science - Southern Arkansas University (2023-01 - 2024-08)