Summary
Senior ML/AI engineer with 7 years' experience building and shipping
production systems. Spent the last three years focused on LLM-powered
product features — RAG pipelines, NL-to-SQL, semantic search — in
Series B and C SaaS environments. Strong Python background,
comfortable owning systems end-to-end from architecture through
on-call. Looking for a role where AI is genuinely core to the product,
not a bolt-on.
Experience
Kairon Analytics · Series C SaaS, 180 people, London
-
Led the build of Kairon's natural-language query feature — an
NL-to-SQL pipeline serving 3,000+ end users across 85 enterprise
customers. Reduced average query resolution time by 62% versus the
previous manual analyst workflow.
-
Designed and owns the RAG layer: vector embeddings of customer
schema, sample queries, and business glossary ingested via
LangChain, stored in pgvector on RDS. Context window management
handles schemas of up to 400 tables without degradation.
-
Built the LLM evaluation framework: automated regression testing
across 1,200+ query/answer pairs with weekly human-in-the-loop
review. Cut hallucination rate from 8.4% to 1.1% over 9 months.
-
Migrated the inference layer from GPT-3.5 to a hybrid GPT-4o /
Claude 3.5 Sonnet setup based on cost-accuracy trade-off analysis,
reducing inference cost by 34% with no accuracy regression.
-
Mentors two mid-level ML engineers; co-authored the team's LLM
engineering guidelines.
Sortly Technologies · Series A e-commerce SaaS, London
-
Built demand forecasting models (XGBoost, Prophet) integrated into
Sortly's inventory management product; forecasting now used by 40%
of the active customer base.
-
Owned the feature engineering pipeline (PySpark, AWS Glue),
reducing data prep time from 6 hours to 45 minutes per training
run.
-
Introduced model monitoring via Evidently AI; set up drift
detection and alerting that caught two significant distribution
shifts before customer impact.
-
Collaborated with product on the insight notification system —
ML-detected anomalies surfaced as plain-English alerts to
customers.
Retail Insight Group · Data consultancy, London
-
Delivered ML engagements for retail clients (churn prediction,
basket analysis, price elasticity modelling) in Python and R.
-
Built internal NLP pipeline for customer feedback classification
using spaCy and a fine-tuned BERT classifier.
Skills
LLM / GenAI
RAG pipelines, NL-to-SQL, prompt engineering, function calling, LLM
evaluation, LangChain, LlamaIndex, OpenAI API, Anthropic API
ML / Python
PyTorch, Scikit-learn, XGBoost, Pandas, Numpy, FastAPI,
PySpark
Vector / Search
pgvector, Pinecone, OpenSearch, text-embedding-3-large
Infrastructure
AWS (ECS, Lambda, RDS, S3, Glue), Terraform, Docker, GitHub
Actions
Data
PostgreSQL, dbt, Snowflake, Redshift
Education
MEng Computer Science (Machine Learning)
University of Edinburgh · First Class Honours
2013 – 2017
Back to the shortlistFICTIONAL SAMPLE CV