Skip to content

TJ Chalokia

AI Engineer | LLM Systems | Data Engineering

Liverpool, England | tejasbirsingh@gmail.com | linkedin.com/in/--tj

LLM-focused engineer building production classification, triage automation, and RAG systems. Former NHS data engineer with 2.5 years shipping production pipelines ahead of schedule, 60% efficiency gains, and a 1st Class MEng with an ML signal-processing thesis.

Available immediately.

Experience

AI Engineer, Indie Consultant

2024 - Present

  • Built LLM triage system using unlabelled CRM data, replacing manual review with bounded expected-value classification
  • Designed feature extraction and hierarchical classification flows validated against real CRM conversion outcomes
  • Built production RAG pipeline with dual encoders, multimodal retrieval, pgvector, FastAPI/Celery, and hallucination checks
  • Built time-series forecasting pipeline with distribution-shift handling for operational intelligence

Data Engineer, NHS England

2022 - 2024

  • Owned Python, PySpark, and SQL pipelines for Adult Social Care and Cancer Registration, shipping milestones 2 weeks early
  • Built statistical data-quality monitoring and KPI tracking, improving operational efficiency by 60%
  • Led product, engineering, and stakeholder delivery, reducing project delivery time by 20%
  • Designed operational intelligence dashboards for real-time data-driven decision making

Data Analyst / Associate Data Engineer, NHS Digital

2021 - 2022

  • Co-developed data-quality maturity indexing system using SQL, DBT, and orchestration, improving efficiency by 60%
  • Optimized Databricks transformation pipelines with Python, SparkSQL, and CI/CD deployment

Selected Systems

LLM Triage System - production-ready classification for commercial lead review

  • Multi-category classification from natural-language CRM applications
  • Bounded expected-value framework for conversion, time savings, and profitability

Production RAG System - dual-encoder retrieval for accuracy-critical applications

  • Multimodal retrieval with metadata filtering and chunking optimization
  • Retrieval evaluation, source attribution, confidence scoring, and hallucination diagnosis

Technical Focus

  • LLM/RAG: dual encoders, multimodal retrieval, chunking strategies, hallucination diagnosis, evals
  • ML systems: classification, triage automation, expected-value optimization, NLP, forecasting
  • Infrastructure: Python, SQL, PySpark, FastAPI, Celery, PostgreSQL/pgvector, Redis, Databricks, DBT

Education

University of Leeds - MEng, BEng Chemical and Materials Engineering, 1st Class Honours

2017 - 2021

  • ML thesis: PCA for signal processing and time-series trend extraction in nanoparticle synthesis optimization
  • 2018 Dean's List Award for Academic Excellence

Contact

Email or LinkedIn for AI engineering, data systems, and applied LLM work.