Minesh Gandhi

AI Researcher

Bengaluru, Karnataka, India8 yrs 9 mos experience

Most Likely To SwitchAI Enabled

Key Highlights

Expert in deploying deep learning models and GenAI solutions.
Proven track record in driving digital transformation with Azure.
Strong problem-solving skills with multi-project management.

Stackforce AI infers this person is a Data Science and AI specialist with expertise in cloud technologies and automation.

Contact

Skills

Core Skills

Artificial Intelligence (ai)Large Language Models (llm)AutomationNatural Language Processing (nlp)Document ProcessingData TransformationData EngineeringTime Series AnalysisBusiness IntelligenceComputer VisionCloud MigrationMachine LearningContent Development

Other Skills

PyTorchBERTLangGraphClaude 3.5OCRGPT-4SQLpandaspdfminerpython-docxStreamlitGenAIRAGPython (Programming Language)Extract, Transform, Load (ETL)

About

- Senior Data Scientist with extensive experience in creating and deploying deep learning models and GenAI solutions, including agentic AI frameworks using LangGraph to address complex automation and reasoning tasks. - Proven track record in driving digital transformation via cloud platform technologies and data modernization using Azure. - Strong problem-solving and analytics skills with the ability to handle multiple projects concurrently. - Expertise in data science and GenAI skills including Machine Learning, NLP, Deep Learning, Forecasting, Spark ML, LLMs, LangGraph-based multi-agent orchestration, and large-scale AI-driven automation. - Experience designing data and information architecture for Business Intelligence and decision support systems, including data warehouses, metadata management, ETL, data staging, and cloud migration strategies. - Expertise in Azure Data Modernization using Azure Data Lake, Azure Data Factory, Azure Synapse, Databricks, Power BI, Logic Apps, Log Analytics, and related cloud services.

Experience

8 yrs 9 mos

Total Experience

2 yrs 2 mos

Average Tenure

2 yrs 8 mos

Current Experience

Ibm

Senior Data Scientist - Artificial Intelligence

Oct 2023 – Present · 2 yrs 8 mos · Bengaluru, Karnataka, India

Designed and delivered multi-agent frameworks integrating Large Language Models (LLMs), classical ML (BERT), and tools (OCR, DB queries, AWS services) to automate complex workflows across healthcare, finance, taxation, and government domains.
Developed LLM-as-a-Judge systems and benchmarking pipelines to evaluate intent detection accuracy and compare LLM and BERT model outputs.
Built natural language to SQL and pandas code converters using GPT-4, enabling users to query large datasets without coding.
Delivered end-to-end document processing solutions for PDF/Word highlighting and XML transformations, combining LLM reasoning with tools like pdfminer, python-docx, and Streamlit. Applied Retrieval-Augmented Generation (RAG) and semantic search (Milvus) to enhance prompt engineering and improve outputs from models like LLaMA 3.3-70B, LLaMA 2-70B, MPT-7B, Claude 3.5, and CodeLlama 70B. Orchestrated GenAI-powered decision logic for automation of claim approvals, invoice matching, and compliance validation.
Key projects include:
Multi-agent intent detection framework using LLaMA 3.3-70B, BERT, and LLM-as-a-Judge evaluation.
Claim and purchase workflow automation using LangGraph, Claude 3.5, OCR, DB, and rule validation.
Natural language to SQL/pandas code generation with GPT-4 for querying large and small datasets.
Dynamic document highlighting for PDFs and Word files with integrated UI support.
XML file transformation and validation using GenAI + RAG-driven code.
GenAI knowledge augmentation with domain embeddings and semantic search.
Automated XML-to-CSV conversion with CodeLlama-70B, improving accuracy and reducing manual effort.

Artificial Intelligence (AI)PyTorchLarge Language Models (LLM)

Dimensional strategies inc.

Senior Data Engineer

Jun 2022 – Sep 2023 · 1 yr 3 mos · India · Remote

Designed and developed end-to-end data ingestion and transformation pipelines using Azure Data Factory, Azure Databricks, Synapse, and related technologies.
Delivered data cleansing, validation, transformation, and curation across multiple projects, supporting analytics and reporting needs for large datasets.
Created detailed design documents, operations runbooks, unit test cases, and deployment plans to ensure smooth transitions across development, staging, and production environments.
Built predictive models for time series forecasting and designed Power BI reporting layers for critical business metrics.
Key projects include:
Ingested and cleansed bus route data, deployed to curated layer using Azure Data Factory and Databricks. Designed a time series model (Spark ML, Darts library) to predict bus trip ridership with R² of 0.78.
Developed data pipelines to ingest and transform revenue and hours data from Dynamics 365, SQL databases, and SharePoint Excel files. Used Logic Apps for SharePoint ingestion, validated data with Great Expectations, and logged results in Log Analytics. Built Power BI models for revenue, milestone bonus, and hours reporting.
Created data pipelines to ingest Banner data via Synapse pipelines, deployed to Cleanse layer, validated with Great Expectations, and transformed to Curate layer using complex SQL procedures and union logic. Delivered unit tests, documentation, and a cutover plan for production deployment.

Python (Programming Language)Extract, Transform, Load (ETL)Data EngineeringData Transformation

Ibm

Data Scientist - Artificial Intelligence

Jul 2018 – Jun 2022 · 3 yrs 11 mos · Bengaluru, Karnataka, India

Delivered data engineering, computer vision, and automation solutions across domains including utilities, automotive, and finance.
Designed and implemented scalable pipelines, machine learning models, and automation systems using cloud and open-source technologies.
Contributed to code quality, architecture, and deployment best practices while ensuring business logic accuracy and operational efficiency.
Key projects include:
Configured Azure components, designed data models, and developed ingestion pipelines between on-prem Oracle, Azure Data Lake, Azure SQL Database, and Synapse using Azure Data Factory and Databricks. Migrated PL/SQL procedures into PySpark scripts and integrated pipelines with UI.
Built computer vision models for car model detection using YOLO, SSD, and VGG16 with transfer learning. Achieved mean average precision (mAP) of 85% on Indian car datasets. Designed a real-time inference pipeline processing up to 15 FPS for applications like traffic surveillance and parking management.
Created a machine learning model to automate email classification, replacing regex-based logic and improving recall from 60% to 94%. Enhanced invoice and PO number extraction modules through additional training, threshold tuning, and preprocessing, contributing to end-to-end automation.

Python (Programming Language)SQLData EngineeringComputer Vision

Indian institute of technology, kharagpur

Teaching Assistant

Jun 2017 – May 2018 · 11 mos

Served as Teaching Assistant for Programming and Data Structures course. Responsibilities involved mentoring the students, ensuring that the labs were productive for the students, and helping the professor in overall smooth running of the course.

Ibm india private limited

Extreme Blue Internship

May 2017 – Jul 2017 · 2 mos · Bengaluru, Karnataka, India

Cognitive Content Library (Learner-Aware Content) :
Developed learner-aware content and recommendation system to promote personalization, provide immediate feedback to users, mastery map of users to fill in gaps in explanation and lecture materials.
To make video lecture content cognitive, engaging and learner aware using NLP techniques.
Key features- intelligent interaction and instant evaluation with feedback.
Technology used: Watson speech to text, Watson Natural Language Understanding (NLU), NLP techniques for text preprocessing, similarity, scoring and summarization.