Sankalp Chourasia

AI Researcher

Noida, Uttar Pradesh, India5 yrs 10 mos experience

Most Likely To SwitchAI ML Practitioner

Key Highlights

Expert in AI and NLP solutions for healthcare.
Achieved 96% accuracy in LLM-based chatbot.
Skilled in deploying scalable semantic search solutions.

Stackforce AI infers this person is a Healthcare Data Scientist specializing in AI and NLP technologies.

Contact

Skills

Core Skills

Natural Language Processing (nlp)Generative AiSemantic SearchAws Lambda

Other Skills

AI AgentsAmazon RedshiftAmazon Simple Notification Service (SNS)Amazon Web Services (AWS)Analytical SkillsArtificial Intelligence (AI)Computer ScienceContainerizationContinuous Integration and Continuous Delivery (CI/CD)Data ScienceDatabase DesignDeep LearningDockerFastAPIGPT-3

About

As a Senior Data Scientist at Innovaccer, I develop AI and NLP solutions to drive innovation in healthcare. I’m passionate about using AI to improve healthcare outcomes. I'm always exploring new technologies to push the boundaries of what's possible in applied NLP.

Experience

5 yrs 10 mos

Total Experience

1 yr 3 mos

Average Tenure

3 yrs 10 mos

Current Experience

Innovaccer

4 roles

Senior Data Scientist

Promoted

Jan 2024 – Present · 2 yrs 5 mos · Noida

Leading the migration of app architecture to agentic workflows using LangGraph and MCP for enhanced clinical actionability.
Spearheaded shift to LLM based Text2SQL chatbot for US healthcare data using Claude Sonnet achieving 96% benchmark accuracy.
Designed unified and normalized views in Snowflake DB over complex data schemas to streamline the data model and establish structured guidelines for LLM integration.
Built a backend module for efficient management and retrieval of conversation history and feedback along with prompt and app configurations using MongoDB.
Integrated Milvus Vector DB with bge-m3 flag encoders as a scalable semantic search solution for efficient information retrieval of high dimensional clinical proper noun search spaces.
Redesigned and deployed this system on an asynchronous FastAPI based backend with Redis & Celery improving TAT by 40%.
Managed CI/CD-driven product deployments using Helm on AWS and Azure, leveraging Docker and Kubernetes.

Generative AILarge Language Models (LLM)PyTorchDockerSnowflakeMongoDB+6

Data Scientist

Jul 2022 – Dec 2023 · 1 yr 5 mos · Noida

Built a Text to SQL Chatbot using spaCy & Stanza achieving over 60% accuracy on benchmark dataset
Fine tuned a spaCy3 transformer with synthetic data enhancing numerical range detection accuracy by 35%.
Synchronous Backend API endpoints built using FastAPI.
Developed a semantic search pipeline using Sentence Transformers leveraging both bi encoders and cross encoders for efficient retrieval and ranking of relevant information.
Leveraged this pipeline to build an in house chatbot for real time RCA of product tickets by integrating with Slack and Confluence APIs.
Productionized the solution via SageMaker Serverless and AWS Lambda, cutting potential infrastructure costs by 80%.

PyTorchGitAmazon RedshiftNatural Language Processing (NLP)PostgreSQLNatural Language Understanding+7

Data Science Intern

Jan 2022 – Jun 2022 · 5 mos · Noida

Built OCR modules for unstructured clinical text extraction from scratch using Tesseract and PyMuPDF
Created Post OCR correction NLP modules and integrated them with existing coding workflows.
Automated Clinical NER workflows with SciBERT & UMLS increasing downstream clinical term capturing.
Created a crosswalk utility for over 200 biomedical source vocabularies using UMLS.

Predictive AnalyticsGitSDK developmentNatural Language Processing (NLP)Python (Programming Language)spaCy