Headline

AI Engineer @ Kiksy.ai | Building Production AI Systems | Generative AI | Voice AI | AI Agents | RAG | Python

About

Hi, I'm Preet Panchal, an AI Engineer passionate about building production-ready AI systems that solve real business problems. Over the past year, I've worked across Machine Learning, Generative AI, Voice AI, and Conversational AI, developing intelligent applications that automate customer engagement, support, and enterprise workflows. I enjoy turning AI concepts into scalable products that people can actually use. Currently, I'm working as an AI Engineer, where I build production-grade AI solutions, including Voice AI agents, Retrieval-Augmented Generation (RAG) systems, and LLM-powered applications. Previously, during my internship, I developed an ML-based anomaly detection system for Zero Trust Network Access (ZTNA), focusing on user behavior analytics and real-time threat detection. What I work with • Generative AI & Large Language Models (LLMs) • Voice AI & Conversational AI • AI Agents & Agentic Workflows • Retrieval-Augmented Generation (RAG) • Machine Learning & Deep Learning • FastAPI & Python Backend Development • Prompt Engineering & AI System Design Technologies Python • FastAPI • LangChain • LangGraph • TensorFlow • Scikit-learn • PostgreSQL • Docker • Git • Vector Databases I'm particularly interested in designing scalable AI architectures that combine LLMs, tool calling, knowledge retrieval, and backend systems to deliver reliable, production-ready AI experiences. I'm always excited to connect with engineers, founders, and AI enthusiasts to discuss Generative AI, Voice AI, machine learning, or interesting product ideas. Feel free to reach out if you'd like to collaborate or exchange ideas.

Location

Greater Bengaluru Area

AI stack

claude code

cursor

elevenlabs

gemini

hugging face

langchain

Work experience

Artificial Intelligence Engineer at Kiksy.ai, Dec 2025 - Present · 10 mos, Greater Bengaluru Area, Working on production-grade Voice AI, Conversational AI, and Generative AI solutions for enterprise customers, building scalable AI systems that automate customer support, sales, and business workflows. Key Contributions: • Designed and deployed LLM-powered voice and chat AI agents for enterprise clients, automating customer engagement, lead qualification, appointment booking, and support workflows while handling 1,000+ daily interactions. • Built low-latency conversational AI pipelines using FastAPI, integrating ASR, LLMs, TTS, and telephony platforms to enable real-time voice conversations with sub-second response times. • Developed Retrieval-Augmented Generation (RAG) systems by building knowledge ingestion pipelines, document processing workflows, vector search, and prompt orchestration to deliver context-aware AI responses. • Engineered AI agent workflows with prompt engineering, tool calling, conversation state management, and multi-step reasoning to improve response quality, reliability, and task completion. • Processed and analyzed large-scale conversational datasets to optimize intent recognition, prompt performance, and overall user experience, improving intent classification accuracy by approximately 30%. • Integrated multiple LLM providers and continuously evaluated model performance to optimize latency, response quality, and operational cost across production deployments. • Collaborated closely with product, engineering, and customer teams to deliver scalable AI solutions for real-world enterprise use cases. Technologies: Python, FastAPI, Generative AI, Voice AI, LLMs, RAG, LangChain, LangGraph, AI Agents, Prompt Engineering, Vector Databases, PostgreSQL, Docker, Git

Intern at InstaSafe, Jul 2025 - Nov 2025 · 5 mos, Bengaluru, As an AI Intern, I worked on machine learning solutions for Zero Trust Network Access (ZTNA) products, focusing on user behavior analytics and backend optimization. Key Contributions: • Designed and developed an anomaly detection system to identify suspicious user login activities using machine learning models, achieving approximately 92% detection accuracy. • Built a real-time inference pipeline for classifying login events and generating alerts for potential security threats. • Performed extensive feature engineering on authentication logs, including login frequency, IP behavior, device information, session patterns, and temporal features, reducing false positives by approximately 25%. • Evaluated and optimized multiple anomaly detection models to improve detection robustness and production reliability. • Contributed to backend optimization during the migration of services from Java to Go, improving API response times, reducing compute overhead, and enhancing system scalability. • Collaborated with engineering teams to integrate ML models into production workflows for real-time security monitoring.

Education

Bachelor of Computer Applications, Computer Science, Oxford college of science

Public links

github: https://github.com/PreMob