ARJIT SHUKLA 22411007

I am a generative AI engineer with hands-on experience building real-time voice agents, RAG pipelines, and high-performance inference engines. I combine strong software engineering with machine learning to deliver production-ready systems.

At Curelink I engineered a real-time customer support voice agent using AssemblyAI, LangChain, and Cartesia, optimizing end-to-end latency below 700ms through asynchronous generators and producer-consumer patterns. I integrated Claude-Haiku with persistent conversation memory and custom tool calling to automate intents and reduce operational overhead.

My projects include a PDF-based conversational RAG pipeline with HuggingFace embeddings and persistent memory, a two-stage DCGAN face generator with progressive training for super-resolution, and a credit card risk model achieving 97.0% ROC-AUC using ensemble methods and SMOTE. I also engineered a C++ inference engine optimized with Eigen, SIMD, OpenMP, and 8-bit quantization for edge deployment.

I lead creative teams as Joint Secretary of a 20-member production group, managing budgets and resources and winning top awards at Inter-IIT cultural meets. I seek roles where I can build scalable GenAI products, optimize ML systems for performance, and collaborate across engineering and research teams.