Siddharth Ahuja

Real Time Medical Scribe AI:

• Led stakeholder discussions for feature prioritization in Real Time Medical Scribe AI.
• Evaluated and selected optimal speech-to-text APIs.
• Oversaw data format design and UI/UX enhancements.
• Developed a comprehensive API for transcript summarization using ChatGPT, BARD, and other models.
• Implemented efficient load balancing of API with AWS Lambda.
• Integrated real-time status updates into the database.

Medical Scribe AI:

• Evaluated speech-to-text services for batch transcriptions optimization.
• Trained segmentation, categorization, and summarization models to detect topics, assign categories, and generate medical reports using a template engine.
• Developed an AWS pipeline with S3 buckets, SQS queues, and Lambda for efficient report generation from audio uploads.
• Tested various LLMs to enhance report quality.
• Leveraged ChatGPT to create more concise and informative summaries.
• Employed a combination of preprocessing techniques, prompt tuning, and post-processing techniques to enhance and optimize the report generation process.

Dashboard Migration and Data Update Automation:

• •Supervised migration of a dashboard from Excel to Power BI, set up SQL views and schedule updates.

Knowledge Extraction System

• Prioritized scraped paragraphs for intended entities.
• Extracted entity specific datafields with DistilBERT for Question Answering.
• Implemented post processing techniques to remove noise.
• Increased Data Asset production by 500%.

Similar News Clustering System

• Grouping news documents using Cosine Similarity on document vectors.
• Decreased document storage requirement by 73%.

Text and Link Relevance Scorer

• Executed SQL queries to retrieve essential data.
• Developed a Text and Link relevance scoring classification model.
• Enhanced team productivity by 66%.

Periodic Data Asset Development Analytics

• Automated report generation and delivery for data asset metrics tracking.
• Enhanced productivity by 50% and reduced error rates by 83%, leveraging Azure Blob Storage for storage and email delivery.

Public Power BI Dashboards

• The Coronavirus Ticker
• Visualizing India’s Padma Awards

Language Translation System

• Implemented a cost-effective translation system using Chrome and Selenium, reducing expenses by 99%.

ETL Pipeline for Power BI Dashboard:

• Engineered a framework for converting JSON dumps into dataframes, facilitating the creation of a mirrored client database. Integrated daily updates and S3 backup for seamless connection to an analytics dashboard.

Hiring Framework and Data Analysis:

• Automated hiring data processes and improved efficiency.
• Developed scripts for email automation, deck integration, and process enhancement.
• Identified and addressed recruitment bottlenecks.
• Provided actionable insights to stakeholders.