Pushpendra Singh Dahiya

I am an AI Developer with over 10 years of experience in building and deploying intelligent systems across Natural Language Processing (NLP), Speech, and Computer Vision domains. I hold an M.Tech in Computer Technology from IIT Delhi, where my thesis focused on robotic manipulation using computer vision—laying the foundation for my deep interest in AI-driven automation and perception.

Over the years, I’ve had the opportunity to lead and contribute to multiple cutting-edge projects that combine practical AI with business value. I specialize in building end-to-end Generative AI solutions, particularly LLM-powered applications, RAG (Retrieval-Augmented Generation) based chatbots, and agentic workflows that dynamically interact with tools, users, and knowledge bases to produce intelligent outcomes.

At Philips, I led the development and deployment of a production-grade RAG-based chatbot for the Pregnancy+ app. The system uses GPT-based LLMs, LangChain, and a vector graph database to generate personalized, medically grounded responses to user queries. This project significantly improved user engagement and retention and was recognized internally for its innovation. I also drove the LLMOps practices for post-deployment monitoring, evaluation, and compliance to ensure robust performance and user safety.

Prior to this, at Xcalibrate, I led the AI efforts for an executive coaching platform, developing multiple NLP components including semantic search, emotion detection, question generation, and an open-ended introspective conversation bot to support personal transformation journeys.

My core technical skills include PyTorch, TensorFlow, OpenCV, LangChain, Docker, and Kubernetes, and I’m proficient in building scalable systems that integrate seamlessly into production environments. I also have a strong background in classical ML, having built invoice parsing systems using Graph Attention Networks and speech-to-text engines for low-resource languages.

I am passionate about solving real-world problems through intelligent systems that are not only smart but also reliable, interpretable, and user-centric. My approach combines deep technical knowledge with a strong product mindset, allowing me to translate abstract ideas into impactful AI solutions.