Work

Datavalley Inc

Junior AI/ML Engineer

Sep 2025 – Present · Remote · EdTech & AI SaaS

Production AI and ML engineering inside the Datavalley CRM — a proprietary SaaS platform used by internal and client marketing, sales, and operations teams.

Outcomes:

  • → ~12% higher lead-to-opportunity conversion from behavioural lead scoring model deployed in production
  • → 15+ automation pipelines orchestrating 30+ services and components
  • → Agentic LangChain workflows replacing manual decision steps across lead qualification and campaign optimisation
  • → LLM-powered RAG pipelines with advanced prompting (CoT, ReAct) for intelligent content and decision layers
  • → REST APIs exposing model predictions for real-time use in CRM UI and internal tooling
  • → Trained and mentored 100+ learners on Python, ML, and MLOps

Stack:

Python·FastAPI·LangChain·LlamaIndex·N8n·GCP (Vertex AI, Cloud Run, Cloud Storage)·MLflow·Docker·MongoDB·NeonDB·Qdrant·Supabase

Academic Projects

University of Roehampton, MSc Data Science AI/ML

Enterprise Knowledge RAG Assistant

Hybrid retrieval (semantic + keyword) with grounded answer generation and citations.

Results: 92% answer relevance · 87% factual accuracy · sub-second query latency

Stack: Python · LangChain · Pinecone · OpenAI API

OCR Resume Pipeline

Automated resume extraction pipeline processing 2,000+ resumes using GCP Cloud Functions and Gemini API into structured JSON.

Stack: Python · GCP Cloud Functions · Gemini API · Cloud Storage

Moving Object Detection (Published)

CNN-based pipeline for real-time object detection supporting attendance capture and emotion recognition.

Results: 94% accuracy (attendance) · 89% accuracy (emotion)

Stack: PyTorch · Computer Vision