Engineering Leader & Principal Architect · Gurugram, India
Engineering leader and architect with 20+ years building and scaling enterprise and SaaS platforms — from cloud-native architecture to AI-enabled delivery. Most recently Associate Director of Software Engineering at Morningstar, leading a 60-person engineering team.
Open to Director of Engineering, VP Engineering, and Senior Architect roles — based in Gurugram / Delhi NCR, open to remote work and to relocation. Also available for select independent advisory engagements through AdviseSphere.
Track record
Career highlights from senior engineering and architecture roles held prior to founding AdviseSphere — not AdviseSphere client outcomes. See the full career history →
About
Dhananjay Singh, Engineering Leader & Principal Architect · Founder, AdviseSphere
Connect on LinkedIn
Dhananjay Singh is a Gurugram-based engineering leader and principal architect, an alumnus of IIT Roorkee, with more than 20 years of experience building and scaling enterprise and SaaS platforms. He has led a 60-person engineering team in India.
His background includes cloud-native architecture, platform modernisation, application modernisation, large-scale cloud migration, observability implementation, and DevOps transformation for global systems.
He has led engineering organisations through architecture design, delivery oversight, production reliability, and AI-enabled software delivery, with experience across India, Australia, the UK, the US, and Europe.
Career experience spans senior engineering and architecture roles at Morningstar, Tata Consultancy Services, Genesys International, and RMSI — including platform delivery for organizations such as Royal Mail, H&M, Roche, Starbucks, and government programmes in India, Wales, and Mongolia.
AdviseSphere products
CRMPro and PortfolioPro are AdviseSphere's own live AI-enabled products for financial advisers.
Hands-on AI fluency
Leading AI-enabled delivery well means understanding what's underneath. I've built the core mechanics myself, from raw tensor operations through a working RAG pipeline, so architecture and investment decisions are grounded in how these systems actually work, not just how to call them.
Implemented multi-head self-attention and a full Transformer block from raw tensor operations — Query/Key/Value projections, scaled dot-product attention, and positional encoding — without relying on pre-built attention layers.
Trained models end-to-end using a custom-built training loop (autograd, loss functions, optimizers), including deliberately inducing and diagnosing an exploding-gradient failure.
Fine-tuned a pretrained language model (GPT-2) and diagnosed overfitting directly from generated output, including near-verbatim memorization.
Built and evaluated a working RAG pipeline: semantic search via cosine similarity, document chunking strategy, and retrieval accuracy evaluation.
Practiced system-design reasoning for scaling a RAG pipeline from a handful of documents to 10 million+ documents at production query volumes.
Worked through the full LLM lifecycle conceptually and hands-on: tokenization, pretraining objective, fine-tuning vs. RLHF, and applied agent/tool-calling patterns.
This reflects self-directed technical training undertaken independently, not enterprise or production work performed for an employer or client.
Advisory services · AdviseSphere
Architecture review sprints, AI readiness audits, legacy modernisation, and fractional technical advisory for teams that want an experienced outside perspective.
Let’s talk
Recruiters and hiring managers: I'm open to Director of Engineering, VP Engineering, and Senior Architect roles — Gurugram / Delhi NCR, remote, or relocation.
For architecture, modernisation, or AI advisory, see AdviseSphere advisory services. For CRMPro or PortfolioPro, see the products page.