Thusharkanth

Founder & Tech Lead

Nexus Solutions • Autônomo

Jan 2026 - Present • 6 mos

I am the Founder and Tech Lead at Nexus Solutions, the technology division of THE MATRICES group. I lead technical strategy, software architecture, and the development of scalable digital solutions, including custom software, SaaS platforms, AI systems, automation tools, and enterprise web applications. My responsibilities include designing system architectures, defining technical roadmaps, and overseeing end-to-end product delivery. I work closely with full-stack developers, UI/UX designers, AI engineers, and interns to build reliable, production-ready solutions for businesses and organizations. My expertise spans full-stack development, AI/LLM systems, SaaS architecture, business automation, backend engineering, cloud deployment, and data engineering. I have hands-on experience with RAG systems, LangChain, LangGraph, vector databases, API orchestration, and cloud-native applications. Previously, I worked as an AI/ML Engineer Intern at MintHRM, where I contributed to an AI-powered HR chatbot using RAG architecture, multi-agent workflows, vector databases, and LLM integrations. I am passionate about solving complex technical challenges and building scalable, maintainable software that delivers measurable business value. At Nexus Solutions, we continue to advance AI-driven systems, enterprise software, and digital transformation solutions for clients worldwide.

AI/ML Engineer Intern

Infosys • Período integral

Sep 2025 - Feb 2026 • 5 mos

Worked as an AI/ML Engineer Intern at MintHRM, contributing to the design and development of an AI-powered HR assistant system built on Retrieval-Augmented Generation (RAG) architecture. Responsible for building document ingestion pipelines, chunking strategies, embedding generation, and vector database integration to enable semantic search and intelligent retrieval. Developed and optimized a LangGraph-based multi-agent system for structured task execution, including supervisor-agent orchestration, tool routing, and workflow control. Implemented LLM integration using open-source models (LLaMA 3, Qwen via Ollama), along with prompt engineering and tool-calling mechanisms to improve response accuracy and system reliability. Worked on backend API development, containerized services using Docker, and integrated OCR, speech-to-text, and text-to-speech modules for multimodal AI capabilities. Collaborated on system architecture design, testing, and internal technical documentation for production readiness and scalability.