G M Kavya

Software Engineer

kavyakavuias@gmail.com+916362787529Bengaluru, IN

Summary

Software Engineer around 2 years of Professional experience and AI/GenAI experience through continuous technical upskilling and end-to-end project development. Experienced in building LLM-powered applications, Agentic AI systems, tool-calling agents, RAG pipelines, and multi-step AI workflows using Python, Gemini, LangChain, FAISS, and Flask. Strong understanding of LLM orchestration, tool selection and execution, embeddings, semantic search, NLP, REST APIs, and production-style deployment. Built and deployed practical AI solutions with modular architectures designed for scalability and maintainability.

Experience 4 yrs 7 mos

Accolite Digitals, Bengaluru, IN
Software EngineerApril 2022 - July 2023 (1 yr 4 mos)
  • Rebuilt and maintained major sections of Accolite's corporate website using HTML, CSS, Bootstrap, and JavaScript as part of a 4-member Agile development team.
  • Translated business requirements and UI designs into responsive web features, owning implementation, debugging, and pre-release validation of assigned modules.
  • Collaborated with designers and backend developers across Agile sprints to troubleshoot issues, integrate changes, and deliver features on schedule.

Skills: HTML5, CSS3, JavaScript, ReactJS, Bootstrap, Responsive Web Development, Debugging & Pre-release Validation, Agile/Scrum, Git/Version Control

Independent Technical Upskilling & AI Project Development, Bengaluru, IN
Self-Directed AI/ML EngineerAugust 2023 - Present (3 yrs 3 mos)
  • 2+ years of hands-on experience independently building depth in Python and applied AI/GenAI — including LLM applications, RAG, AI agents, tool calling, NLP, APIs, and deployment — through multiple end-to-end software projects.
  • Built and deployed practical AI systems that integrate LLMs with retrieval pipelines, external/internal tools, structured context, and multi-step workflows, using modular, API-based architectures with clear separation between retrieval, tool, and LLM layers for scalability and maintainability.
  • Applied engineering practices across the full development lifecycle, including API design, validation, persistence, frontend integration, testing/debugging, and cloud deployment.

Skills: Python, LLM Application Development, Prompt Engineering, Google Gemini API, LangChain, Ollama (local LLMs), Retrieval-Augmented Generation (RAG), FAISS, Embeddings, Semantic Search, AI Agents, Tool-Calling, Multi-Step Workflow Orchestration, Flask, REST API Design

Education

BMS Institute of Technology, B.E., Computer Science
2018 - 2022

Bengaluru, IN, Grade: 8.54/10

Skills

Tool-Calling Agents, Multi-Step Tool Execution, Agent Design, LLM Orchestration, Prompt Engineering, Google Gemini API, LangChain, Retrieval-Augmented Generation (RAG), FAISS, Embeddings, Semantic Search, Ollama (local LLMs), Python, Flask, REST APIs, SQLAlchemy, Pydantic, Authentication, SQL Server, MySQL, Pandas, Scikit-learn, spaCy, TF-IDF, Named Entity Recognition (NER), ReactJS, JavaScript, HTML5, CSS3, Bootstrap, Git, GitHub, Docker, Vercel, Render, Agile/Scrum, T-SQL, Window Functions, Self-Joins, Statistical Anomaly Detection, ETL Pipeline Design, Data Validation, SQLite

Projects

AI Email Intelligence & Automation Platform github.com/kavya608/ai-email-intelligence-platform
  • ● Built and deployed an AI-powered email automation platform using Python, Flask, React, Gemini, and NLP to classify emails, summarize messages, identify priorities, extract entities, action items, and deadlines.
  • ● Implemented a Retrieval-Augmented Generation (RAG) pipeline using Gemini embeddings and FAISS to retrieve relevant company-policy context and generate grounded responses.
  • ● Developed a Gemini-powered tool-calling AI agent capable of selecting and executing email-search and policy-retrieval tools to answer multi-step queries using retrieved context.
  • ● Designed REST APIs and integrated the AI backend with a React dashboard; deployed the application using Render and Vercel.
  • ● Architected the system with modular, decoupled layers for retrieval, tool execution, and LLM orchestration, enabling scalable extension to new tools and data sources.
Chat with PDF
  • ● Built a local RAG application for natural-language Q&A over PDF documents using LangChain for orchestration, PyPDFLoader and recursive text splitting for document processing, and FAISS for vector search.
  • ● Ran inference fully locally via Ollama (Qwen2.5:1.5B) using langchain-ollama, with HuggingFace sentence-transformer embeddings (all-MiniLM-L6-v2) — an end-to-end RAG pipeline independent of hosted LLM APIs.

Languages

  • English, Professional
  • Kannada, Native