Vanshika Uttarwar profile photo 1 on Tal

Vanshika Uttarwar

Data Scientist | Generative AI | LLMs | AI Agents | Python | LangGraph | FastAPI | AWS | Healthcare AI

Bengaluru

5+ years of work experience

ScribeEMR Systems Private Limited
Aug 20261 moPresent

Designed and deployed enterprise-grade AI and Generative AI solutions for healthcare documentation automation using Python, FastAPI, LangChain, LangGraph, OpenAI, Gemini, Claude, PostgreSQL, and AWS. * Built intelligent document understanding pipelines that transformed unstructured clinical guidelines into structured prompts and JSON schemas, reducing manual configuration effort by over 80%. * Developed and optimized multi-LLM orchestration frameworks with dynamic model selection, fallback strategies, retry mechanisms, and concurrency controls to improve reliability, latency, and cost efficiency. * Evaluated and optimized LLM performance through prompt engineering, experimentation, and output validation, delivering production-ready solutions for 100+ healthcare providers with improved extraction accuracy and reduced hallucinations. * Designed multimodal AI pipelines using Vision Language Models to extract structured information from handwritten and scanned clinical documents, achieving 93–96% extraction accuracy while reducing inference costs by 52%. * Built AI-powered speech intelligence pipelines using Whisper and Large Language Models to generate structured clinical documentation, extracting 25+ medical entities with over 98% accuracy. * Analyzed model performance, monitored production metrics, and implemented continuous improvements to enhance accuracy, robustness, scalability, and operational efficiency. * Developed scalable backend services and data pipelines using FastAPI, PostgreSQL, asynchronous processing, RBAC, and GitLab CI/CD to support production AI systems. * Built agentic AI workflows using LangGraph, MCP, and N8N to automate reporting, notifications, and business processes, reducing backend processing latency by 95%. * Partnered with clinicians, product managers, and cross-functional stakeholders to translate business requirements into AI-driven solutions aligned with healthcare documentation standards.
ScribeEMR Systems Private Limited
Aug 20251 yr 1 moAug 2026

Designed and developed enterprise-grade Generative AI applications for healthcare documentation automation using Python, FastAPI, LangChain, LangGraph, OpenAI, Gemini, Claude, PostgreSQL, and AWS. • Built AI workflows that automatically transformed clinical guideline documents into structured prompts and JSON schemas, reducing manual configuration effort by over 80%. • Engineered reliable multi-LLM orchestration pipelines with intelligent model routing, fallback strategies, retry mechanisms, and concurrency controls for production workloads. • Developed production-ready prompt engineering frameworks for 100+ healthcare providers, improving extraction accuracy while minimizing hallucinations and ensuring consistent outputs. • Designed multimodal AI pipelines to extract structured information from handwritten and scanned clinical documents using Vision Language Models, achieving 93–96% extraction accuracy while reducing inference costs by 52%. • Built AI-powered speech-to-documentation workflows using Whisper and LLMs to generate structured clinical notes, extracting 25+ medical attributes with 98%+ accuracy. • Developed secure backend services with FastAPI, PostgreSQL, RBAC, asynchronous processing, and GitLab CI/ CD, enabling scalable deployment of AI services. • Built agentic AI workflows using LangGraph, MCP, and N8N to automate internal reporting, notification systems, and business workflows, reducing backend latency by 95%. • Collaborated with clinicians, product managers, operations teams, and business stakeholders to design AI solutions aligned with healthcare documentation standards.
Schoolknot
Aug 202411 mosJun 2025

Remote

• Developed and optimized data reporting solutions for a 1M+ record ERP platform, improving reporting accuracy by 30% and enabling data-driven decision-making. • Performed exploratory data analysis to uncover product adoption trends and user engagement gaps, providing insights that informed FY25 product strategy. • Built executive Tableau dashboards for fee recovery, attendance, and operational KPIs, helping improve billing efficiency by 10%. • Designed and launched an AI-powered mock examination platform that delivered personalized performance analysis and automated student feedback using LLMs. • Created intelligent assessment workflows to evaluate student responses, identify learning gaps, and generate targeted improvement recommendations.
Encora Inc.
Jul 20222 yrs 2 mosAug 2024

Pune District

• Developed responsive web applications for healthcare and ed-tech clients using ReactJS, JavaScript, HTML, and CSS, delivering intuitive and scalable user experiences. • Built Python and SQL-based analytics solutions that enabled large-scale reporting, operational visibility, and data-driven business decisions. • Automated reporting and data-processing workflows using Pandas, saving 40+ hours of manual effort per month and improving reporting efficiency. • Collaborated with cross-functional teams on API integrations, database design, testing, and Agile delivery to ensure high-quality software releases. • Created reusable UI components and implemented unit testing best practices, improving application maintainability, consistency, and overall code quality.
Encora Inc.
Jan 20226 mosJun 2022

Pune

Encora Inc.
Jun 20217 mosDec 2021

Pune, Maharashtra, India

GitHub
@vanshikauttarwar1
Medium
@vanshikauttarwar14
Personal portfoliods-portfolio-sepia.vercel.app

Vanshika’s AI Stack

antigravity
antigravity
claude code
claude code
elevenlabs
elevenlabs
grok
grok
hermes agent
hermes agent
hugging face
hugging face

Education

Full Stack Data Science & AI

AlmaBetter (Accredited by IIT Guwahati)

20252025
Smt.Kashibai Navale college of Engineering

Bachelor of Engineering, Information Technology

Smt.Kashibai Navale college of Engineering

20174 yrs2021

Join Vanshika and 10k+ other professionals on

Tal Boss

Tal is the app where bosses meet engineers & builders directly, without middlemen

Download Tal Boss

Vanshika Uttarwar on Tal

Headline

Data Scientist | Generative AI | LLMs | AI Agents | Python | LangGraph | FastAPI | AWS | Healthcare AI

About

Hello! I’m vanshika, Completed my engineering from Pune university. I am an AI Engineer with experience building AI-powered applications, intelligent automation systems, and scalable backend services. My work focuses on Large Language Models (LLMs), AI Agents, Agentic Workflows, Retrieval-Augmented Generation (RAG), and production-ready AI systems using Python, FastAPI, LangGraph, AWS, and PostgreSQL. I began my career as a Frontend Developer, building responsive web applications with ReactJS and modern JavaScript frameworks. Over time, I expanded into Data Analytics and AI Engineering, combining software engineering fundamentals with machine learning and generative AI technologies. I enjoy designing and developing AI solutions that automate workflows, enhance decision-making, and create measurable business impact. My interests include Agentic AI, Multi-Agent Systems, LLM Orchestration, AI Product Development, and scalable AI infrastructure. I am currently seeking opportunities to build next-generation AI products and intelligent systems that solve real-world business problems.

Location

Bengaluru

AI stack

antigravity

claude code

elevenlabs

grok

hermes agent

hugging face

Work experience

Data Scientist at ScribeEMR Systems Private Limited, Aug 2026 - Present · 1 mo, Designed and deployed enterprise-grade AI and Generative AI solutions for healthcare documentation automation using Python, FastAPI, LangChain, LangGraph, OpenAI, Gemini, Claude, PostgreSQL, and AWS. * Built intelligent document understanding pipelines that transformed unstructured clinical guidelines into structured prompts and JSON schemas, reducing manual configuration effort by over 80%. * Developed and optimized multi-LLM orchestration frameworks with dynamic model selection, fallback strategies, retry mechanisms, and concurrency controls to improve reliability, latency, and cost efficiency. * Evaluated and optimized LLM performance through prompt engineering, experimentation, and output validation, delivering production-ready solutions for 100+ healthcare providers with improved extraction accuracy and reduced hallucinations. * Designed multimodal AI pipelines using Vision Language Models to extract structured information from handwritten and scanned clinical documents, achieving 93–96% extraction accuracy while reducing inference costs by 52%. * Built AI-powered speech intelligence pipelines using Whisper and Large Language Models to generate structured clinical documentation, extracting 25+ medical entities with over 98% accuracy. * Analyzed model performance, monitored production metrics, and implemented continuous improvements to enhance accuracy, robustness, scalability, and operational efficiency. * Developed scalable backend services and data pipelines using FastAPI, PostgreSQL, asynchronous processing, RBAC, and GitLab CI/CD to support production AI systems. * Built agentic AI workflows using LangGraph, MCP, and N8N to automate reporting, notifications, and business processes, reducing backend processing latency by 95%. * Partnered with clinicians, product managers, and cross-functional stakeholders to translate business requirements into AI-driven solutions aligned with healthcare documentation standards.

Gen AI Engineer at ScribeEMR Systems Private Limited, Aug 2025 - Aug 2026 · 1 yr 1 mo, Designed and developed enterprise-grade Generative AI applications for healthcare documentation automation using Python, FastAPI, LangChain, LangGraph, OpenAI, Gemini, Claude, PostgreSQL, and AWS. • Built AI workflows that automatically transformed clinical guideline documents into structured prompts and JSON schemas, reducing manual configuration effort by over 80%. • Engineered reliable multi-LLM orchestration pipelines with intelligent model routing, fallback strategies, retry mechanisms, and concurrency controls for production workloads. • Developed production-ready prompt engineering frameworks for 100+ healthcare providers, improving extraction accuracy while minimizing hallucinations and ensuring consistent outputs. • Designed multimodal AI pipelines to extract structured information from handwritten and scanned clinical documents using Vision Language Models, achieving 93–96% extraction accuracy while reducing inference costs by 52%. • Built AI-powered speech-to-documentation workflows using Whisper and LLMs to generate structured clinical notes, extracting 25+ medical attributes with 98%+ accuracy. • Developed secure backend services with FastAPI, PostgreSQL, RBAC, asynchronous processing, and GitLab CI/ CD, enabling scalable deployment of AI services. • Built agentic AI workflows using LangGraph, MCP, and N8N to automate internal reporting, notification systems, and business workflows, reducing backend latency by 95%. • Collaborated with clinicians, product managers, operations teams, and business stakeholders to design AI solutions aligned with healthcare documentation standards.

Data Analyst at Schoolknot, Aug 2024 - Jun 2025 · 11 mos, Remote, • Developed and optimized data reporting solutions for a 1M+ record ERP platform, improving reporting accuracy by 30% and enabling data-driven decision-making. • Performed exploratory data analysis to uncover product adoption trends and user engagement gaps, providing insights that informed FY25 product strategy. • Built executive Tableau dashboards for fee recovery, attendance, and operational KPIs, helping improve billing efficiency by 10%. • Designed and launched an AI-powered mock examination platform that delivered personalized performance analysis and automated student feedback using LLMs. • Created intelligent assessment workflows to evaluate student responses, identify learning gaps, and generate targeted improvement recommendations.

Software Developer at Encora Inc., Jul 2022 - Aug 2024 · 2 yrs 2 mos, Pune District, • Developed responsive web applications for healthcare and ed-tech clients using ReactJS, JavaScript, HTML, and CSS, delivering intuitive and scalable user experiences. • Built Python and SQL-based analytics solutions that enabled large-scale reporting, operational visibility, and data-driven business decisions. • Automated reporting and data-processing workflows using Pandas, saving 40+ hours of manual effort per month and improving reporting efficiency. • Collaborated with cross-functional teams on API integrations, database design, testing, and Agile delivery to ensure high-quality software releases. • Created reusable UI components and implemented unit testing best practices, improving application maintainability, consistency, and overall code quality.

Junior Software Engineer at Encora Inc., Jan 2022 - Jun 2022 · 6 mos, Pune

Trainee software Engineer at Encora Inc., Jun 2021 - Dec 2021 · 7 mos, Pune, Maharashtra, India

Education

Full Stack Data Science & AI, AlmaBetter (Accredited by IIT Guwahati)

Bachelor of Engineering, Information Technology, Smt.Kashibai Navale college of Engineering

Public links

github: https://github.com/vanshikauttarwar1

medium: https://medium.com/@vanshikauttarwar14

ds-portfolio-sepia.vercel.app: https://ds-portfolio-sepia.vercel.app/