Open to opportunities

Vedansha Srivastava

@vedanshasrivastava

Message

I build production AI agents, RAG systems, and full-stack applications that automate user support.

India

Message

What I'm looking for

I'm looking to build AI, machine learning, and full-stack systems that solve real user problems, improve through data-driven decisions, and prioritize performance optimization.

At AssessHub, I build and deploy an AI-powered candidate support agent used by clients including HDFC Bank and ICICI Lombard across WhatsApp and email.

I engineered a PostgreSQL, pgvector, Gemini embeddings, and LLM-reranking RAG pipeline that routes 4,000+ candidate interactions across 40+ support intents. I also built feedback loops, regression-checked knowledge-base updates, and an operations dashboard with RBAC, compliance logging, and human escalation queues.

I’ve also built a real-time agentic voice assistant with wake-word activation, streaming voice responses, document retrieval, browser automation, Auth0 authentication, and MongoDB session memory.

For audio authenticity, I created a dual-layer watermarking and verification system using Wav2Vec2, Librosa, PyTorch, and NumPy, delivered through a Streamlit app for real-time scoring. I’m focused on solving user problems through AI systems, data-driven decisions, and performance optimization.

Experience

Work history, roles, and key accomplishments

Current

Software Development Engineer Intern

Current

Assesshub

May 2026 - Present (3 months)

Built and deployed an AI-powered candidate support agent used in production by clients including HDFC Bank and ICICI Lombard, automating support across WhatsApp and email. Engineered a RAG pipeline with PostgreSQL, pgvector, and Gemini embeddings to route 4,000+ candidate interactions across 40+ support intents, and developed a self-improving feedback system with an ops dashboard.