Professional experience

I turn analysis into working AI systems.

My career spans insurance, legal services, higher education, AI research, and sustainability. I began by making complex data useful, then moved into building the products, infrastructure, and adoption paths that put AI into practice.

Now: End-to-end AI and automation delivery at ERM, alongside data and AI consultancy for Oxford.
Before: Applied data science at Google DeepMind and enterprise analytics leadership at Mishcon de Reya.
Grounding: Insurance analytics at Marsh and more than three years of part-time teaching at UCL.

Career timeline

Roles and impact

Current and overlapping roles appear first. Open any role for the fuller technical and operational breakdown.

Jan 2026 – Present London, UK Current

ERM

AI Solutions Manager

Sole end-to-end owner of ERM's AI & Automation applications: product design, full-stack development, cloud infrastructure, CI/CD, and adoption. Translate frontline workflows into production AI, choosing build, buy, or configure across custom Streamlit, Claude, Copilot, and n8n so colleagues can use the right tool for the job.

Delivery, platform, and governance details

Shipped in-house

6 AI and automation tools live at ERM URLs.
Two-agent system for ERM's Impact Assessment AI using Docling, clustering, and approximately 100 per-section skills.
Few-shot document extraction with source verification.
Agentic requirements-gathering flow, calculation QA engine, and threaded scraper.

Platform and governance

Provisioned and owned AWS CDK, ECS Fargate, ALB, and CloudFront infrastructure.
Own ERM's n8n Cloud instance as a governed low-code layer.
Git-based source control through GitHub, SSO, and audit logging to IT and security standards.
Led demos, roadshows, and L2 support, collapsing processes from months to minutes.

Full-stack AI
AWS CDK / ECS / ALB / CloudFront
Agentic systems
n8n
Streamlit
Claude / Copilot
CI/CD

Sep 2023 – Present Remote, UK 2+ years · Current

University of Oxford

Consultant, Data & AI

Co-created CLARA, the Corporate Litigation Research Assistant: an autonomous agent that supports corporate litigation and accountability research by searching a knowledge base of more than 1.2 million pages of litigation-relevant archival documents.

CLARA infrastructure and data pipeline details

Infrastructure and deployment

Automated AWS-based deployments using ECR, ECS, and Lambda.
GitHub Actions CI/CD pipeline.
Streamlit app for usage statistics monitoring.

Data pipeline

SageMaker for data processing.
Dagster for data pipeline orchestration.
ETL for chatbot data analysis.

AWS
Docker
CI/CD
GraphRAG
LLMs
ECS
ECR
Lambda
DynamoDB
CloudWatch
Route 53

May 2025 – Oct 2025 London, UK 6 months · Core Analytics Group

Google DeepMind

Applied Data Scientist

Part of the Core Analytics Group, applying data analytics and AI to drive insights across people, culture, and organisational health.

Responsibilities and internal tools

Key responsibilities

Worked with people, talent, performance, and evaluation data.
Built secure, permission-aware data tables to support reporting.
Built an MVP of 2 specialised AI agents with AI Studio.

PLX ecosystem and tools

PLX Scripts.
PLX Workflows.
GoogleSQL and F1 Query.

PLX Scripts
PLX Workflows
PLX Dashboards
GoogleSQL
F1 Query

Feb 2022 – May 2025 London, UK 3+ years · Lead Analytics Advisor

Mishcon de Reya

Principal Data Scientist

Served as the firm's lead data science and analytics advisor, engaging with C-suite leadership to drive data-informed decision-making. Delivered the firm's first business operations KPI report by capturing key measures across all functions.

70% reduction in report redundancy

10+ functions with certified data models

20% self-serve adoption

Transformation, leadership, and cross-functional details

Cloud infrastructure and data migration

Led migration from SSRS to Power BI.
Led migration from SQL Server to Azure Synapse.
Established a single source of truth.
Reduced report redundancy by 70%.

Data democratisation and self-serve

Led the Data Center of Excellence.
Implemented self-serve analytics.
Increased adoption of centralised reporting.
Integrated data across departments.

Technical leadership

C-suite stakeholder engagement.
Data strategy development.
Line management and mentoring.
Vendor and API integration.

Analytics and insights

Certified data models across 10+ functions.
Client segmentation using RFM.
Time forecasting and machine learning exploration.
Data governance automation.

People and HR

Talent, performance, and evaluation data.
Attrition and headcount analytics.
Gallagher system integration.

Finance and legal

Billing, debtors, and cash receipts.
Legal matters data.
Matter partner analytics.

Business development

Client segmentation and retention.
Marketing analytics framework.
HubSpot and GA4 integration.

Power BI
Azure Synapse
SQL Server
Azure Data Factory
Power Apps
Power Automate

Oct 2021 – Feb 2025 London, UK 3+ years · Part-time

UCL School of Management

Teaching Assistant

Supported teaching and marking in MSc Business Analytics modules at UCL School of Management, working with academic module leads through small-group teaching and coursework assessment.

Teaching responsibilities and modules

Teaching responsibilities

Seminar and workshop facilitation.
Weekly office hours.
Coursework assessment and marking.
Student feedback provision.

Module support

Marketing Analytics.
Data Engineering.
Programming with Python.
Predictive Analytics.

Marketing Analytics
Data Engineering
Python
Predictive Analytics

Jun 2021 – Nov 2021 London, UK 6 months · Insurance Analytics

Marsh

Data Analyst

Focused on claims analysis and produced insights reports requested by clients after analysing data received from several insurance companies. Automated manual analysis processes using Python tools.

Achievements and technologies

Key achievements

Reduced report generation time by 50%.
Automated manual analysis processes.
Manipulated large insurance datasets.
Generated client insights reports.

Technologies used

Microsoft Excel, including VLOOKUP, pivot tables, and macros.
Python using Pandas and PyExcel.
Data manipulation and cause-mapping.
Cross-insurance-company data analysis.

Python
Pandas
Excel
Data Analysis

Contact

Building something difficult with AI, automation, or data?

Email is the simplest place to start. A short note about the context and what is currently stuck is enough.

Email Meghana Connect on LinkedIn

Higher education · Oct 2021–Feb 2025

Teaching Assistant

UCL School of Management · MSc Business Analytics

Supported workshops, office hours and assessment across Python, data engineering, predictive analytics and marketing analytics.