Hi! I'm Sumantha NTS
AI/ML Enthusiast based in India
I am currently working as a Senior Machine Learning Engineer with 4+ years experience in handling data science, machine learning, computer vision and Generative AI projects.
I am currently working as a Senior Machine Learning Engineer with 4+ years experience in handling data science, machine learning, computer vision and Generative AI projects.
sumanthneechadi@gmail.com
Bengaluru, Karnataka
+91 9035697661
B.E in Electrical & Electronics
Spearheaded the development of a client-facing Finetuning as a Service (FTAAS) platform, enabling the customization and deployment of open-source LLMs using LoRA/QLoRA and KubeFlow.
Acted as a key liaison between business units and technical teams, translating complex problems into actionable AI-driven solutions.
Mentored and guided junior Machine Learning Engineers, fostering their professional development and maximizing team contribution.
Led development of a text-to-SQL MVP and a conversational chatbot for data-driven insights.
Fine-tuned multiple LLMs (BERT, DeepSeek) for specialized legal domain tasks, achieving high accuracy.
Engineered and deployed an agentic workflow for automated report generation using CrewAI, FastAPI, and Qdrant.
Directed an object detection project using Yolo-v8 to identify vehicle defects with 92% mAP.
Developed several LLM-powered tools to automate tasks like slide generation and resume parsing.
Conducted advanced experiments with generative image models, including fine-tuning Stable Diffusion and implementing ControlNet.
Built and deployed a high-accuracy inventory forecasting system using XGBoost to predict stock levels.
Analyzed industrial power system networks, executing detailed power flow studies and delivering comprehensive technical reports to optimize grid performance and ensure reliability.
Comprehensive PDF manipulation tool featuring Merge, Split, Cut, Rotate and an AI-powered Chat with PDF capability using RAG.
Generated synthetic data using various strategies with large language models (LLMs) and fine-tuned a BERT model for classification tasks in the legal domain.
Created a process which automatically generates a overview for a given contract which include complete summary, responsibilities, risk associated with the parties.
Developed an intelligent chatbot that analyzes and provides insights on financial data, helping users make informed decisions.
Smart resume parser using YOLO v5 for section detection and NER models to extract structured information from resumes.
Comprehensive image editing tools including background removal, upscaling, style transfer, and text-to-image generation with an intuitive interface.
In Retrieval-Augmented Generation (RAG), the quality of your embedding model dictates the relevance of retrieved context. A superior embedding model captures semantic nuances better than keyword matching, directly impacting the LLM's ability to generate accurate and grounded responses.
Read ArticleRecent benchmarks show that reasoning models (like o1 or specialized CoT models) significantly outperform standard LLMs on complex tasks requiring multi-step logic. While non-reasoning models excel at pattern matching, reasoning capabilities are essential for autonomous agents and complex problem-solving.
Read ArticleBuilding reliable AI Agents requires a shift from prompt engineering to system engineering. Key guidelines include: defining clear tools, implementing robust error handling loops, ensuring state management, and designing for observability to debug agent thought processes effectively.
Read ArticleI'm always open to discussing new projects, creative ideas, or opportunities to be part of your visions.
For project inquiries
sumanthneechadi@gmail.comLet's connect professionally
View ProfileCheck out my code
View RepositoriesBangalore, India
Open to Opportunities