GenAI Engineer & Quantum Machine Learning Specialist
Designed and delivered production AI systems spanning
enterprise Generative AI, RAG, agentic AI, LLM infrastructure,
and quantum machine learning and optimisation.
Served as a technical lead and go-to expert for GenAI,
RAG, agentic architectures, and LLM system design.
April 2020 – August 2026
Enterprise GenAI
NICO – Enterprise Generative AI & RAG Platform
Architected Nokia's production enterprise GenAI platform,
consolidating fragmented technical knowledge into
natural-language search, question answering, summarisation,
reasoning, and tool-enabled workflows for
1,000+ global daily users
across 1M+ documents.
Owned key architecture decisions across document ingestion,
parsing, chunking, embeddings, vector indexing, retrieval,
and RAG orchestration. Designed a service architecture
separating backend and RAG/LLM components, with Celery
and Redis for asynchronous processing and Qdrant for
large-scale vector retrieval.
Agentic AI
Multi-Agent Architecture & Enterprise Tooling
Designed and developed the agentic AI architecture
from the ground up using LangGraph,
MCP, and a dynamic
Tool Registry, including planner,
tool selector, executor, reviewer/reviser,
and final-answer agents.
Enabled enterprise workflows spanning Jira and Confluence
search, deep research, Excel calculation and extraction,
and presentation generation through context-aware
and permission-aware tool selection.
Retrieval & Security
Large-Scale Retrieval, Reranking & Access Control
Designed permission-aware retrieval using subgroup
definitions, collection-level restrictions, and
retrieval-time access checks to prevent cross-group
information leakage in enterprise RAG workflows.
Evaluated Qwen, Llama, Mixtral, and GPT-family models
against quality, latency, and deployment requirements,
and introduced LLM-based reranking
to improve context selection at million-document scale.
AI Infrastructure
LLM Inference, Optimisation & Observability
Defined GPU and Triton-based inference architecture,
KV caching, and inference optimisation strategies
for production LLM workloads.
Applied caching and ReWOO-style planning to reduce
redundant LLM calls and repeated context processing.
Established AI observability through
MLflow,
Grafana,
and production monitoring across multiple environments.
Technical Leadership
Architecture, Engineering Leadership & Mentoring
Guided engineering implementation across backend,
frontend, DevOps/platform, domain, and support teams,
while acting as a technical lead for GenAI, RAG,
agentic AI, LLM architecture, and optimisation.
Delivered global technical presentations,
knowledge-sharing sessions, and mentoring to
engineers and managers.
