Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

Hi, I'm Kishan Ajudiya—a dedicated Applied AI & MLOps Engineer with 3+ years building production AI systems for enterprise environments. I've deployed and served open-source LLMs (LLaMA, Qwen, Mistral) with OpenAI-compatible APIs on Kubernetes using vLLM inference, handling ~500 concurrent users with on-prem data privacy. I have a passion for turning complex data into reliable AI-driven outcomes and helping teams operationalize AI safely and at scale. I have a proven track record delivering enterprise-grade RAG systems—from chunking and embeddings to retrieval, generation and LLM-as-judge evaluation—reducing manual work and operational overhead. I specialize in Python, Agentic AI, cloud infrastructure, microservices, and system design, with a focus on security, compliance, and robust observability.…Hi, I'm Kishan Ajudiya—a dedicated Applied AI & MLOps Engineer with 3+ years building production AI systems for enterprise environments. I've deployed and served open-source LLMs (LLaMA, Qwen, Mistral) with OpenAI-compatible APIs on Kubernetes using vLLM inference, handling ~500 concurrent users with on-prem data privacy. I have a passion for turning complex data into reliable AI-driven outcomes and helping teams operationalize AI safely and at scale. I have a proven track record delivering enterprise-grade RAG systems—from chunking and embeddings to retrieval, generation and LLM-as-judge evaluation—reducing manual work and operational overhead. I specialize in Python, Agentic AI, cloud infrastructure, microservices, and system design, with a focus on security, compliance, and robust observability.

Kishan Ajudiya

AI Engineer, Data Scientist, Data Analyst, +2





Hi, I'm Kishan Ajudiya—a dedicated Applied AI & MLOps Engineer with 3+ years building production AI systems for enterprise environments. I've deployed and served open-source LLMs (LLaMA, Qwen, Mistral) with OpenAI-compatible APIs on Kubernetes using vLLM inference, handling ~500 concurrent users with on-prem data privacy. I have a passion for turning complex data into reliable AI-driven outcomes and helping teams operationalize AI safely and at scale. I have a proven track record delivering enterprise-grade RAG systems—from chunking and embeddings to retrieval, generation and LLM-as-judge evaluation—reducing manual work and operational overhead. I specialize in Python, Agentic AI, cloud infrastructure, microservices, and system design, with a focus on security, compliance, and robust observability.…Hi, I'm Kishan Ajudiya—a dedicated Applied AI & MLOps Engineer with 3+ years building production AI systems for enterprise environments. I've deployed and served open-source LLMs (LLaMA, Qwen, Mistral) with OpenAI-compatible APIs on Kubernetes using vLLM inference, handling ~500 concurrent users with on-prem data privacy. I have a passion for turning complex data into reliable AI-driven outcomes and helping teams operationalize AI safely and at scale. I have a proven track record delivering enterprise-grade RAG systems—from chunking and embeddings to retrieval, generation and LLM-as-judge evaluation—reducing manual work and operational overhead. I specialize in Python, Agentic AI, cloud infrastructure, microservices, and system design, with a focus on security, compliance, and robust observability.

Available to hire

Hi, I’m Kishan Ajudiya—a dedicated Applied AI & MLOps Engineer with 3+ years building production AI systems for enterprise environments. I’ve deployed and served open-source LLMs (LLaMA, Qwen, Mistral) with OpenAI-compatible APIs on Kubernetes using vLLM inference, handling ~500 concurrent users with on-prem data privacy. I have a passion for turning complex data into reliable AI-driven outcomes and helping teams operationalize AI safely and at scale.

I have a proven track record delivering enterprise-grade RAG systems—from chunking and embeddings to retrieval, generation and LLM-as-judge evaluation—reducing manual work and operational overhead. I specialize in Python, Agentic AI, cloud infrastructure, microservices, and system design, with a focus on security, compliance, and robust observability.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Language

English

Fluent

Work Experience

Applied AI & MLOps Engineer at Hochschule Flensburg

January 1, 2025 - Present

Designed and shipped end-to-end production agentic RAG assistant—from chunking strategy and retrieval tuning to generation and LLM-as-judge evaluation—cut manual document lookup time by 84%. Built a modular ingestion→embedding→retrieval→generation→evaluation pipeline for rapid model experimentation. Developed a DeepEval-based LLM-as-judge evaluation with MLflow tracing for reproducibility. Deployed open-source LLMs and vision-language models on bare-metal Kubernetes using vLLM for ~500 concurrent users with GDPR on-premises data privacy. Engineered a GPU VRAM capacity planner to guide hardware procurement. Added auth, rate-limiting, logging and tracing; Prometheus/Grafana dashboards provided live cluster visibility. Fine-tuned YOLOv11 with SAM 2-assisted annotation achieving 92.54% accuracy.

Computer Vision Engineer at AL-KO Geräte GmbH

April 1, 2024 - September 1, 2024

Developed a real-time obstacle avoidance system for a robotic lawn mower using YOLO instance segmentation; achieved 90.2% mAP and 87.9% recall on critical object classes. Optimised inference for edge deployment by converting the model to NCNN format on Raspberry Pi 5, reducing latency from 1492ms to 435ms with <1% accuracy drop.

Computer Vision Engineer at Continental AG

January 1, 2023 - December 1, 2023

Filtered redundant sensor data and reformatted it for a Matrix-LED headlight control system, improving dimming response speed for oncoming and preceding vehicles.