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

I am an impact-driven ML Engineer with over 6 years of experience building NLP, Generative AI, and Computer Vision systems. My expertise spans LLM fine-tuning, retrieval-augmented generation, and legal/financial document understanding. I am skilled in building scalable AI pipelines from data preparation to production using Python, PyTorch, Hugging Face, and cloud-native services like Azure and GCP. I have a passion for leveraging machine learning to deliver real-world impact, especially in regulated domains. I excel at designing experiments, integrating models into microservices, and collaborating cross-functionally in agile teams.…I am an impact-driven ML Engineer with over 6 years of experience building NLP, Generative AI, and Computer Vision systems. My expertise spans LLM fine-tuning, retrieval-augmented generation, and legal/financial document understanding. I am skilled in building scalable AI pipelines from data preparation to production using Python, PyTorch, Hugging Face, and cloud-native services like Azure and GCP. I have a passion for leveraging machine learning to deliver real-world impact, especially in regulated domains. I excel at designing experiments, integrating models into microservices, and collaborating cross-functionally in agile teams.

Harshini Gudipally

AI Engineer, Data Scientist, AI Developer





I am an impact-driven ML Engineer with over 6 years of experience building NLP, Generative AI, and Computer Vision systems. My expertise spans LLM fine-tuning, retrieval-augmented generation, and legal/financial document understanding. I am skilled in building scalable AI pipelines from data preparation to production using Python, PyTorch, Hugging Face, and cloud-native services like Azure and GCP. I have a passion for leveraging machine learning to deliver real-world impact, especially in regulated domains. I excel at designing experiments, integrating models into microservices, and collaborating cross-functionally in agile teams.…I am an impact-driven ML Engineer with over 6 years of experience building NLP, Generative AI, and Computer Vision systems. My expertise spans LLM fine-tuning, retrieval-augmented generation, and legal/financial document understanding. I am skilled in building scalable AI pipelines from data preparation to production using Python, PyTorch, Hugging Face, and cloud-native services like Azure and GCP. I have a passion for leveraging machine learning to deliver real-world impact, especially in regulated domains. I excel at designing experiments, integrating models into microservices, and collaborating cross-functionally in agile teams.

Available to hire

I am an impact-driven ML Engineer with over 6 years of experience building NLP, Generative AI, and Computer Vision systems. My expertise spans LLM fine-tuning, retrieval-augmented generation, and legal/financial document understanding. I am skilled in building scalable AI pipelines from data preparation to production using Python, PyTorch, Hugging Face, and cloud-native services like Azure and GCP.

I have a passion for leveraging machine learning to deliver real-world impact, especially in regulated domains. I excel at designing experiments, integrating models into microservices, and collaborating cross-functionally in agile teams.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Microsoft Certified Professional

Intermediate

Work Experience

Senior Gen AI Developer at Manulife (Client)

February 1, 2025 - Present

Developed retrieval-augmented generation (RAG) systems using LangChain, Azure AI Search, and GPT-4 to extract legal clauses from insurance and loan agreements, improving data completeness by 35%. Engineered end-to-end document understanding pipelines combining Azure Document Intelligence, spaCy, and custom models (LayoutLM), reducing annotation errors by 50%. Deployed models in AKS-based microservices with Docker and GitHub Actions CI/CD, aligning with production best practices for availability and observability. Conducted prompt engineering and evaluation of multiple large language models (GPT-4, Claude, Llama) for claims document Q&A workflows, optimizing latency vs cost tradeoffs in multi-model setups. Partnered with legal analysts and software engineers to iterate ML solutions for domain-specific pain points.

Data Scientist at Intellect Design Arena

July 1, 2024 - August 5, 2025

Developed an NLP-based Intelligent Document Processing solution for structured and unstructured document types in the insurance sector using LLM models and custom NER, achieving 98% data extraction precision. Built object detection models using TensorFlow, PyTorch, and transformer-based architectures (DETR, Detectron, YOLO, LayoutLM), boosting data model precision. Designed high-performance classification models for document categorization, achieving accuracy that reduced manual effort by 80% and increased process efficiency. Developed a fraud detection system using graph databases and graph neural networks, achieving 90% accuracy in identifying fraudulent claims. Built custom NER and classification models for legal and insurance document classification using TensorFlow and PyTorch, achieving 98% precision. Implemented A/B tests and experiment tracking (MLflow) to benchmark extraction strategies and document routing logic across product versions.

Associate Solution Leader at Brane Enterprises Private Limited

April 1, 2021 - August 5, 2025

Designed the Natural Solution Language (NSL) project converting BPMN or flowchart diagrams into functional software solutions using generative models and technologies. Deployed AI models on edge devices like NVIDIA Jetson Nano using TensorRT and quantization techniques, increasing processing speed for real-time applications by 50%. Executed the Ergonomics project utilizing Google's pose detection model, analyzing body movements from user videos and providing real-time corrections for improved posture and technique. Created sophisticated algorithms automating extraction of key fields from bank cheques, streamlining banking automation processes and minimizing manual intervention by 90%. Engineered feature extraction and selection processes, reducing dataset complexity and improving model accuracy by 20%.

Machine Learning Engineer at Techolution India Pvt Ltd

August 1, 2019 - August 5, 2025

Created a large-scale facial recognition system for biometric authentication, enhancing security and access control across 15 client locations with high accuracy. Employed image processing techniques with OpenCV and scikit-image, analyzing images and achieving a 50% improvement in clarity and overall image quality. Conducted hyperparameter tuning using Grid Search and Random Search techniques, leading to performance improvements across machine learning models. Developed ensemble models combining decision trees, logistic regression, and neural networks, achieving accuracy improvement for churn prediction.