London, UK
Aditya Thakare
Cloud Architect & Forward Deployed Engineer
12+ years architecting enterprise-scale cloud platforms and AI systems. Recognised expert in GCP, distributed data pipelines, and Agentic AI architectures.
Google Cloud Next
GenAI Platform Architecture Showcase
Architected a self-serve GenAI Conversational Analytics platform that reduced enterprise time-to-insight from weeks to minutes. This architecture was selected and showcased as a flagship enterprise use case at the main Google Cloud Next Keynote stage.
Watch on YouTube (at timestamp 31:10)Experience
Accenture — London, UK
AI/ML Computational Science Manager · Lead Cloud Architect & Hands-on IC
April 2019 — Present
- Architected a self-serve GenAI Conversational Analytics platform enabling 1200+ users to generate SQL and charts via plain English — showcased at the Google Cloud Next Keynote.
- Acted as a direct architectural feedback loop to Google Cloud US Product Engineering, driving upstream fixes and hands-on coding custom middleware solutions.
- Engineered production LLM evaluation pipelines using BigQuery UDFs, targeting 85% SQL generation accuracy with active prompt drift detection.
- Designed a zero-hallucination multi-agent RAG pipeline for MPs querying millions of government documents with strict entailment checks.
- Re-architected a monolithic pipeline into a serverless fan-out system, reducing processing time from 12 hours to 12 minutes across 20 million XML files.
- Led Cloud Engineering for the world's first CDP-to-GCP deployment, reducing data onboarding from 6 weeks to 3 days per 100 tables.
Accenture — Pune, India
Lead Data Engineer · Major UK Telco Account
March 2017 — March 2019
- Architected a "first of its kind" Spark-based Right to Be Forgotten (RTBF) GDPR framework across all Hive tables with automated masking.
- Designed a fault-tolerant REST API interface using complex HBase data models, exposing real-time Smart Cities insights to enterprise clients.
Accenture — India
Software Engineer · UK & Singapore Telecom Clients
2014 — March 2017
- Led complex Data Platform migrations and built customised WebApps for major telecom clients across the UK and Singapore.
Featured Projects
Conversational Analytics Platform
Self-serve platform enabling 1200+ enterprise users to generate SQL queries and interactive charts via plain English. Built with modular Python, TDD, and custom Agent classes.
★ Showcased at Google Cloud Next KeynoteGovAssist — Zero-Hallucination RAG
Multi-agent validation pipeline for MPs querying millions of government documents. Isolated Evaluation Agent performs strict entailment checks to guarantee verifiable citation grounding.
Zero hallucination guarantee60× Pipeline Speedup
Re-architected a monolithic XML processing system into a serverless, dynamic fan-out pipeline using Apache Airflow and Cloud Run. Processed 20 million files, reducing load time from 12 hours to 12 minutes.
12 hours → 12 minutesAgent Builder — One-Click Generation
Autonomous ADK skill that safely explores massive BigQuery schemas via restricted TABLESAMPLE guardrails to auto-generate complete agent configurations, fully integrated with GitLab CI/CD.
Weeks → seconds for agent creationTech Stack
GenAI & ML
Cloud Infrastructure
Data Engineering
DevOps & Security