Agentic HR query resolution
Three cooperating agents — data extraction, knowledge search and voice analytics — answer employees’ HR questions, grounded in the company’s own policy documents through fine-tuning and retrieval-augmented generation.
Agents, vision systems and small language models, and the data pipelines underneath them — taken all the way to the device in someone’s hand. Currently an AI intern at Yitro Global, and finishing a B.Tech in AI & Data Science.
Three cooperating agents — data extraction, knowledge search and voice analytics — answer employees’ HR questions, grounded in the company’s own policy documents through fine-tuning and retrieval-augmented generation.
Wake word, speech-to-text, a local language model that can call tools, and speech synthesis — openWakeWord, faster-whisper, Qwen and Piper wired into one loop that runs entirely on the machine in front of you.
A self-paced learning agent that tailors what it teaches, how fast and how hard to each learner’s progress, grounded with retrieval-augmented generation — and talks the lesson through, using CUDA-accelerated inference and text-to-speech to keep the voice low-latency.
A navigation assistant for visually impaired users: real-time obstacle detection and spoken, hands-free guidance, using object detection and speech recognition on the phone they already carry.
Quantum neural networks, variational quantum classifiers and hybrid classical–quantum models built in Qiskit and TensorFlow, then measured against the classical models a clinician would actually reach for.
A distributed big-data pipeline on HDFS that processes patient prescription data at scale, feeding a machine-learning validation engine that warns patients on several prescriptions about unsafe drug combinations.
AI Intern — AI Algorithms for Enterprise Systems
Engineered an AI tutoring application driven by a small language model, with an interactive agent avatar delivering real-time explanations, quizzes and hands-on coding feedback — alongside a teacher-facing analytics and classroom-management dashboard.
Architected a talent-matching network on graph databases and multi-agent orchestration, helping recruiters find the best candidate-to-role fits through clan-based clustering.
Built an end-to-end CRM platform and designed LLM-driven chatbots that agentize its workflows, automating lead management and customer interactions.
Developed a recruiter–job seeker matching application that uses graph-based clan clustering to improve candidate discovery and engagement.
Designed a multi-agent system — a knowledge-based search agent, a data extraction agent and a voice agent — to automate enterprise information retrieval and query resolution.
I’m Tendool, an AI/ML and data engineer. The part I like best is what happens after the notebook — when a model has to answer a real question, run on real hardware, and be trusted by someone who will never read the code.
Since September 2025 I’ve been an AI intern at Yitro Global, working on small language models, graph-backed multi-agent systems and real-time voice agents for enterprise use. Alongside it I’m finishing a B.Tech in Artificial Intelligence & Data Science at ASE, Coimbatore.
Before that came university projects across quantum ML, healthcare data pipelines, robotics and agriculture. They taught me to care about the unglamorous parts: clean data, honest baselines, and testing against real inputs instead of mocks. I’m based in Bobbili, Andhra Pradesh, and work remotely.
Multi-agent and RAG architectures grounded in an organisation’s own documents, so answers come from policy rather than guesswork.
Real-time detection on small hardware — a phone camera, a Raspberry Pi, a Jetson Nano — with speech in and out.
Quantum neural networks and variational classifiers benchmarked honestly against classical baselines.
Distributed HDFS pipelines that process prescription data at scale, and cleaning and analytics tools tested against the real data before any model is allowed near it.
Predictive models for green-hydrogen electrolysis cells, and crop-disease identification from field photos.
Graph-based clan clustering that surfaces candidate-to-role fits, and LLM chatbots that agentize CRM workflows end to end.
B.Tech, Artificial Intelligence & Data Science
Class XII, MPC — Board of Intermediate Education, AP · Visakhapatnam
Class X — Board of Secondary Education, AP, Bobbili