hi, i'm Lakshmi Anne
senior ai/ml engineer · founder of cwg technology
building production ai systems and personal decision tools.
companies i've built
live products and consultancies serving real customers
open source work
flagship production-grade reference architecture, plus selected projects
enterprise-ai-platform
enterprise ai workflow platform
production reference architecture for multi-tenant rag with tool-using agents
- ▸multi-tenant saas with postgres row-level security
- ▸transactional outbox + kafka for async ingestion
- ▸pgvector hnsw semantic search with citation tracking
- ▸claude sonnet 4.6 agent with custom react loop (no langgraph)
- ▸sse-streamed chat with token, tool_use, citations events
- ▸full observability: opentelemetry, prometheus, loki, tempo, grafana
9 architecture decision records (adr) documenting trade-offs
rag-enterprise-search
→production-grade rag pipeline for enterprise document search
multi-cloud-ai-agent
→ai agent system that orchestrates llms across azure, aws, and gcp
mlops-azure-template
→end-to-end mlops template for ml model deployment on azure
llm-sql-to-python
→llm-powered tool that converts sql queries to python code
sentiment-kafka-pipeline
→real-time sentiment analysis pipeline using kafka and ml models
yolov8-object-tracking
→real-time object detection and tracking using yolov8 and deepsort
production ai at scale
tools i use to ship
7+ years building production ai
i'm a senior ai/ml engineer based in milton keynes, uk. i specialise in shipping ml systems that survive contact with reality — real users, real edge cases, real money on the line.
my work spans regulated industries (banking, hospitality, retail, energy) where ml failures have real consequences. i'm also the founder of cwg technology, an enterprise digital engineering consultancy serving global brands.
let's build something
open to senior ai/ml roles, consulting engagements, and interesting collaborations. best way to reach me is email.
real-time fraud detection
HSBC's existing rules-based fraud system was producing too many false positives, blocking legitimate transactions and frustrating customers. We needed an ML system that could decision in under 100ms at 1M+ tx/day, while reducing false positives without missing real fraud.
Swipe
Stream
Store
+ IF
or Allow
Engine
Score
gpt-4 hospitality platform
Marriott's UK properties were sitting on 200K+ guest reviews scattered across booking platforms, with no scalable way to extract actionable insights. Manually responding to each review took staff hours per day. We needed a GenAI system that could analyse sentiment, surface themes, and draft personalised responses at scale.
Ingest
Pinecone
Retrieval
Reasoning
Dashboard
Extraction
Draft
demand forecasting at scale
Tesco was over-ordering perishables across thousands of stores, leading to high waste and margin loss. The existing forecasting was too coarse — it couldn't account for local weather, regional events, or promotion cannibalisation. We needed an SKU-level forecasting system that could handle 10K+ products across multi-horizon windows.
+ Stock
+ Promo
Pipeline
+ XGBoost
Reorder
Forecast
Stacking
storm risk modelling
Insurance partners needed earlier and more accurate storm impact predictions to prepare claims teams, route field assessors, and warn policyholders. Existing weather models gave general regional warnings but couldn't predict damage severity at postcode level. We needed a model fusing satellite, radar, and historical claims data to score risk 72 hours ahead.
+ Radar
Pipeline
Join
CNN
Alerts
Risk Map
Forecast