Ayub Subhaniya
Gurgaon
Summary
Director of Engineering, ML, with 6+ years leading Research and Infrastructure teams across NLP, Generative AI, Retrieval-Augmented Generation (RAG), and Conversational AI. Developed an in-house LLM for CXM domain achieving 90% GPT-4 performance at 10x costs, improved bot resolution rates from 30% to 60%, and fastened implementation cycles from 4 weeks to under 7 days. Successfully built an engineering team of 11 people with zero attrition for the last 3 years.
Overview
6
6
years of professional experience
Work History
Director of Engineering, ML
Sprinklr
01.2019 - Current
- Customer Care Copilot: Reduced Average Handling Time (AHT) by 16% with a conversational interface integrating fragmented point solutions via agentic workflows.
- FAQ Chatbots: Deployed NLP-driven solutions using Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG), automating query resolution at 40%-60% for global clients.
- Generative AI VoiceBots: Increased automated query resolution from 30% to 50% in low-resource language (Hinglish), focusing on human-like interactions and low latency.
- AI-generated Brand Playbooks: Transitioned from decision-tree bots to text-based using conversation analytics, shortening bot deployment from 4 weeks to under 7 days.
- In-house LLM Development: Developed an internal CXM LLM, achieving ~90% accuracy of GPT-4 at 10x lower costs on domain tasks.
- Customer Care Agent Assistance: Enhanced product quality by over 2x through advanced NLP pipelines, fine-tuned LLMs, and robust RAG systems, adopted by 200+ clients globally.
- Community Search Solution: Designed and implemented a hybrid search solution leveraging Elasticsearch, vector search (dense and sparse retrieval methods), and advanced ranking fusion algorithms, achieving 85% top-3 relevancy with sub-second latency.
- Team Building: Built and scaled a high-performance team to 11 people across research, and ML infrastructure with a sharp focus on technical excellence and ownership. Cultivated a culture of learning, fast iteration, and career acceleration—resulting in 0% attrition and multiple engineers transitioning into senior roles, tech leads, and patent contributors.
Product Engineer Internship
Sprinklr
01.2019 - 04.2019
- Designed and developed a scalable ML inference framework inspired by TensorFlow Serving using Docker and Kubernetes; migrated 50+ NLP models to the new microservice architecture supporting 300+ RPS.
Education
B.Tech - Information and Communication Technology
Dhirubhai Ambani Institute of Information And Communication Technology
Gandhinagar, Gujarat01.2019
Skills
Technical Skills
- Natural Language Processing (NLP)
- Generative AI & LLM
- Retrieval-Augmented Generation (RAG)
- Elasticsearch & Vector Databases
- Model Optimization & Fine-Tuning
- Data Engineering & Pipelines
- Kubernetes & Docker
- Distributed Systems Architecture
- CI/CD
Languages
- Python
- Java
Managing Skills
- Engineering Management & Leadership
- Strategic Innovation & Execution
- Talent Mentorship & Development
- Product Vision & Technical Roadmap
Achievements
- Patent Filed, 2022, Systems and Methods for Performance Advertising Smart Optimizations leveraging Reinforcement Learning.
- Runner-Up, Sprinklr AI Hackathon, 2024, Recognized for innovative NLP and ML solutions.
- You Are a GEM Award, 2024, Honored for outstanding leadership and AI innovation.
Languages
English
Hindi
Timeline
Product Engineer Internship
Sprinklr
01.2019 - 04.2019
Director of Engineering, ML
Sprinklr
01.2019 - Current
B.Tech - Information and Communication Technology
Dhirubhai Ambani Institute of Information And Communication Technology
Similar Profiles
Anoop DevAnoop Dev
Principal Project Manager at SprinklrPrincipal Project Manager at Sprinklr
Durgesh ShuklaDurgesh Shukla
Product Configuration Analyst II at SprinklrProduct Configuration Analyst II at Sprinklr
Kshitij MishraKshitij Mishra
Principal Solution Consultant at SprinklrPrincipal Solution Consultant at Sprinklr