Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

I am Paresh Uchdadiya, an AI / LLM Engineer with 15+ years of IT experience, including 10 years in software engineering and 5 years focused on AI and LLM-driven systems. I design and scale distributed systems, modernize them with LLMs, agentic workflows, enterprise search, and RAG architectures. I have hands-on experience deploying production-grade AI solutions across cloud platforms, prioritizing precision, performance, and reliability in enterprise environments. I excel at building end-to-end pipelines, integrating data infrastructure, and delivering measurable improvements in accuracy and efficiency, from clinical AI triage to secure multi-region service meshes.…I am Paresh Uchdadiya, an AI / LLM Engineer with 15+ years of IT experience, including 10 years in software engineering and 5 years focused on AI and LLM-driven systems. I design and scale distributed systems, modernize them with LLMs, agentic workflows, enterprise search, and RAG architectures. I have hands-on experience deploying production-grade AI solutions across cloud platforms, prioritizing precision, performance, and reliability in enterprise environments. I excel at building end-to-end pipelines, integrating data infrastructure, and delivering measurable improvements in accuracy and efficiency, from clinical AI triage to secure multi-region service meshes.

Paresha Uchdadiya

AI Engineer, Developer, Data Scientist, +2





I am Paresh Uchdadiya, an AI / LLM Engineer with 15+ years of IT experience, including 10 years in software engineering and 5 years focused on AI and LLM-driven systems. I design and scale distributed systems, modernize them with LLMs, agentic workflows, enterprise search, and RAG architectures. I have hands-on experience deploying production-grade AI solutions across cloud platforms, prioritizing precision, performance, and reliability in enterprise environments. I excel at building end-to-end pipelines, integrating data infrastructure, and delivering measurable improvements in accuracy and efficiency, from clinical AI triage to secure multi-region service meshes.…I am Paresh Uchdadiya, an AI / LLM Engineer with 15+ years of IT experience, including 10 years in software engineering and 5 years focused on AI and LLM-driven systems. I design and scale distributed systems, modernize them with LLMs, agentic workflows, enterprise search, and RAG architectures. I have hands-on experience deploying production-grade AI solutions across cloud platforms, prioritizing precision, performance, and reliability in enterprise environments. I excel at building end-to-end pipelines, integrating data infrastructure, and delivering measurable improvements in accuracy and efficiency, from clinical AI triage to secure multi-region service meshes.

Available to hire

I am Paresh Uchdadiya, an AI / LLM Engineer with 15+ years of IT experience, including 10 years in software engineering and 5 years focused on AI and LLM-driven systems. I design and scale distributed systems, modernize them with LLMs, agentic workflows, enterprise search, and RAG architectures.

I have hands-on experience deploying production-grade AI solutions across cloud platforms, prioritizing precision, performance, and reliability in enterprise environments. I excel at building end-to-end pipelines, integrating data infrastructure, and delivering measurable improvements in accuracy and efficiency, from clinical AI triage to secure multi-region service meshes.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Work Experience

AI Engineer at McKesson

March 1, 2023 - Present

Designed and deployed a real-time AI triage assistant using fine-tuned LLMs on clinical documentation, improving diagnostic accuracy by 32% and reducing manual review time by 45% through ReAct-based agentic reasoning. Built autonomous agent workflows with LangGraph to orchestrate symptom extraction, diagnostic reasoning, and report generation enabling self-directed clinical insights with minimal human intervention. Engineered an MCP-based integration layer for real-time multi-source data fusion enhancing LLM-driven triage and ICU alerting. Developed a multimodal diagnostic AI pipeline combining BLIP-2 and GPT-4V for early-stage disease detection. Implemented LangChain RAG with Pinecone for high-precision Q&A over clinical documents, improving triage accuracy by 25%. Automated ingestion pipelines using AWS services to continuously train clinical models and reduce data workload by 35%. Generated synthetic medical images with Stable Diffusion to expand training data, improving recall on r

AI Tools Developer at Infinite Infolab

November 1, 2021 - September 4, 2025

Built scalable ETL pipelines using SQL, Pandas, and NumPy reducing reporting latency by 35%. Developed and deployed ML models for customer churn prediction and segmentation improving retention targeting accuracy by 22%. Automated data workflows increasing pipeline reliability and reducing manual interventions by 40%. Created real-time KPI dashboards with Tableau and Power BI enhancing decision-making for product and marketing teams. Applied NLP for sentiment analysis on support tickets reducing issue escalation by 18%. Built multilingual rule-based chatbots integrated with support systems reducing ticket volume by 30%. Led OCR pipeline development achieving 92% digitization accuracy on handwritten and printed forms. Designed early autonomous chatbot task flows for ticket triaging laying groundwork for agentic AI systems. Developed internal image labeling and annotation tools accelerating dataset preparation by 50%. Conducted A/B testing and statistical analyses for feature rollout eval

LLM Engineer (Contract) at The Home Depot, TX

October 1, 2025 - January 1, 2026

Designed and developed an enterprise search enhancement with recall-first retrieval and precision-first classification, improving top-ranked result quality without increasing latency. Built semantic retrieval pipelines using Vertex AI embeddings, integrating with BigQuery for large-scale attribute extraction and normalization. Implemented query intent classification to route and rank documents, increasing search relevance and user trust. Reduced irrelevant results by 60%, increasing adoption of internal search over manual knowledge discovery. Explored multiple LLMs (Mistral7b, Qwen3) to extract attributes and normalize missing attribute values from BigQuery data. Leveraged sidecar proxies in a service mesh to enable secure, multi-region routing for enterprise search APIs. Conducted BDD, destructive, and load testing to validate API performance, reliability, and scalability. Created and collected metrics to record latency and responses and exposed to Grafana dashboards. Contributed to P

AI Engineer (Contract) at McKesson, TX

March 1, 2023 - September 1, 2025

Designed and deployed a real-time AI triage assistant using fine-tuned LLMs on clinical documentation, improving diagnostic accuracy by 32% and reducing manual review time by 45% via ReAct-based agentic reasoning. Built autonomous LangGraph workflows to orchestrate symptom extraction, diagnostic reasoning, and report generation, enabling self-directed clinical insights with minimal human intervention. Engineered an MCP-based integration layer to streamline access to EHRs, DICOM images, genomic profiles, and real-time vitals for AI pipelines; enhanced performance of LLM-driven triage, imaging analysis, and ICU alerting through context-aware, low-latency data fusion. Implemented a multimodal diagnostic pipeline with BLIP-2 and GPT-4V for cross-modal radiology imaging and clinical notes, enabling early-stage disease detection. Used LangChain RAG with Pinecone for high-precision Q&A over millions of clinical documents, achieving 25% improvement in triage decisions. Automated ingestion of D

AI Tools Developer at Infinite Infolab, India

March 1, 2018 - November 1, 2021

Built scalable ETL pipelines using SQL, Pandas, and NumPy to process customer interactions and business operations data, reducing reporting latency by 35%. Developed ML models (logistic regression, decision trees, random forests) for churn prediction and behavioral segmentation, improving retention targeting accuracy by 22%. Automated data workflows, increasing pipeline reliability and reducing manual intervention by 40%. Created dashboards in Tableau and Power BI for real-time KPI tracking, enabling data-driven decisions across product and marketing. Implemented NLP techniques for sentiment analysis on support tickets, reducing issue escalation by 18%. Built multilingual chatbots and OCR pipelines for document digitization, supporting early autonomous task flows that triaged tasks and routed them to the appropriate departments. Led dataset labeling tooling and validation to support model training and QA.