Thinking in systems
Finding the moving pieces, the failure paths, and the cleanest way to make them cooperate.
Portfolio / experiments
Hi, I am Vrinda
Building things, breaking things, and learning from both.
Current Coordinates
Curious across the stack
I am not trying to fit inside one tiny label. Right now, I am drawn to AI systems, developer tools, cyber security, networks, and software that has to be both clever and dependable.
Making intelligence usable
Exploring agents, tool calling, and workflows that turn models into useful collaborators.
Keeping room to branch out
Following hard problems wherever they go, from browsers to backends to networks.
Selected Work
Projects worth opening
Work Experience
Research intern work
June 2026 - July 2026
CCNCS
Research Intern, Centre for Computer Networks and Cyber Security
Worked with PES University CCNCS on research at the intersection of computer networks, cyber security, LLM evaluation, and model-selection workflows.
Skills
Tools I reach for
Languages
C, C++, Python, JavaScript, Dart
AI / ML
Agentic AI, LLM Agents, ReAct, Tool Calling, Sentence Transformers, Random Forest
Backend & Web
FastAPI, React.js, Node.js, Express.js, WebSockets, CodeMirror
Systems & Networks
Linux, TCP/IP, OpenFlow, Software-Defined Networking, Mininet, Ryu
Data & Tools
SQLite, MongoDB, NumPy, Pandas, Git, GitHub, Grafana, VS Code
Research Experience
Accepted IEEE papers
A Difficulty-Aware Prompt-Model Recommendation Framework for LLM-Generated Python Code Using Competency-Based Evaluation
A research direction focused on evaluating code-generation difficulty and recommending suitable model choices for Python programming tasks.
A Unified Competency-Driven and Difficulty-Aware Framework for Prompt-Model Recommendation in LLM-Based Mathematics
A research framework focused on evaluating mathematical reasoning across competency dimensions and recommending suitable LLM-prompt combinations based on problem difficulty.
About
Curious, technical, and allergic to boring builds.
I like projects where intelligence meets infrastructure: agents that need memory, tools that need good UX, networks that need resilience, and systems that need to hold together when things get messy. I am still exploring where I want to go next, which is honestly the fun part.