Looking for a job

Aayush Thakar

@aayushthakar

Gen AI Engineer at Stealth AI Startup who developed a production conversational voice agent and optimized its latency and cost.

India

Message

At Stealth AI Startup, I developed and iterated a production-deployed conversational AI product for a banking client, combining speech-to-text, LLM reasoning, and text-to-speech. I also developed RAG systems with FAISS, Hugging Face embeddings, and Mistral, and implemented agentic workflows with LangGraph.

At Swiftly PayTech, I built an end-to-end fraud detection solution that reduced false positives by 15%, then deployed it as a REST API with FastAPI and Docker. I hold an M.Sc. in Artificial Intelligence and Machine Learning from University of Limerick and am open to relocating to Singapore.

Experience

Work history, roles, and key accomplishments

Gen AI Engineer

Stealth AI Startup

Jun 2026 - Aug 2026 (2 months)

Worked directly with a banking client to gather requirements for an AI voice-agent solution, translating business needs into technical requirements. Built and iterated a production-deployed conversational AI product integrating speech-to-text, LLM reasoning, and text-to-speech, and developed RAG systems using FAISS, Hugging Face embeddings, and Mistral.

Python RAG Faiss Hugging Face Mistral LangGraph Speech to Text Text To Speech MLFlow CI CD Docker fastAPI AWS SIP

Machine Learning Engineer Intern

Swiftly PayTech

Mar 2024 - Jun 2024 (3 months)

Built an end-to-end fraud detection solution for financial transaction data using Python, XGBoost, and scikit-learn, reducing false positives by 15%. Designed and deployed the model as a REST API using FastAPI and Docker with CI/CD automation.

Python XGBoost scikit learn fastAPI Docker CI CD REST APIs Feature Engineering Fraud Detection

AI/ML Engineer Intern

Zetpeak

Nov 2022 - Jan 2023 (2 months)

Designed automated data preprocessing and feature engineering pipelines for structured datasets and applied statistical analysis to improve model performance. Optimized inference and system monitoring workflows, contributing to a 20% improvement in system uptime.

Python data preprocessing Feature Engineering Statistical Analysis System Monitoring