Kethavath Ajay Kumar

Software Engineer

New Delhi, Delhi, India6 mos experience

AI EnabledAI ML Practitioner

Key Highlights

Expert in machine learning and AI systems development.
Successfully fine-tuned LLaMA 3 for advanced applications.
Developed robust multi-device communication frameworks.

Stackforce AI infers this person is a Machine Learning Engineer with a focus on AI and IoT solutions.

Contact

Skills

Core Skills

Software DesignSoftware IndustryMachine LearningArtificial Intelligence

Other Skills

CUDAMPIOpenMPPythonJavaCC++Dataset CurationPreprocessingFine-TuningLLaMA 3 ModelQuestion-Answer GenerationCode GenerationMQTTMosquitto

About

B.Tech graduate - Computer Science from IIT Delhi, with a strong focus on machine learning and AI systems development. Recently contributed to Shunya AI Research Pvt. Ltd. as an intern, where efforts included fine-tuning a custom LLaMA 3 model for advanced question-answer generation and code generation tasks. Collaborated on dataset curation and preprocessing to enhance model training quality and performance. Proficient in CUDA, MPI, and OpenMP, with hands-on experience building robust frameworks like an MQTT-based multi-device communication system using Mosquitto. Motivated by a problem-solving mindset and a passion for leveraging AI and systems programming to create innovative solutions. Aspiring to contribute to impactful projects in the tech industry while continuing to grow as an engineer.

Experience

Rafay

Software Developer

Feb 2026 – Present · 1 mo · Bengaluru, Karnataka, India · On-site

Software Engineer

Software DesignSoftware IndustryCUDAMPIOpenMPPython+5

Otsi - object technology solutions inc.

Software Engineer

Sep 2025 – Feb 2026 · 5 mos · Hyderabad, Telangana, India · On-site

Software Engineer at OTSi

Software IndustrySoftware Design

Iotiot.in

Software Development Intern

May 2024 – Jul 2024 · 2 mos · Pune, Maharashtra, India · Remote

Internship at Shunya AI Research Pvt. Ltd:
Dataset Curation & Preprocessing: Curated and preprocessed datasets tailored for fine-tuning large language models, significantly improving training quality and model performance.
Training and Fine-Tuning LLMs: Successfully fine-tuned a custom LLaMA 3 model for specialized tasks, focusing on Question-Answer generation and code generation.
Advanced QnA and Code Generation: Designed and implemented a system using the fine-tuned LLaMA 3 model to produce context-aware QnA pairs and accurate, functional code snippets.
MQTT-based Multi-Device Communication Framework: Developed a robust framework using the Mosquitto MQTT broker for seamless real-time communication between server and client devices.
Enabled reliable messaging protocols and confirmation handling across connected devices.
Impact:
Enhanced the efficiency of large language models for domain-specific applications.
Delivered scalable solutions for multi-device communication, demonstrating real-time reliability.