Richard Anthuvan Rosario

Machine Learning & Full-Stack Engineer, Cambridge UK

I'm an engineer working across machine learning and full-stack product development. Right now I build audio data infrastructure and research tooling at Meta Reality Labs in Cambridge.

Before that I was the founding ML engineer at Hearable, getting Transformer audio models running on iPhone, and before that I shipped full-stack products and edge vision systems for a handful of startups.

The easiest way to reach me is on LinkedIn. Some older code lives on GitHub.

Work

  • Audio Software Systems Engineer II, Meta2026 – nowAudio data ingestion, DSP validation and research tooling for Reality Labs Research Audio.

    Meta, Reality Labs Research Audio. Contract, Cambridge, UK. Jun 2026 – Present.

    • Build and maintain the audio data ingestion pipeline that supplies AI research, taking raw device recordings through to validated datasets for model development.
    • Built an internal data-collection platform with React and FastAPI, used by researchers across multiple studies, and own its deployment and DevOps.
    • Implement audio algorithms and DSP processing in the pipeline, with automated validation for clipping, silence, sync and metadata before recordings reach research datasets.
    • Sole systems engineer on the team, owning the network, management and lifecycle of the prototype audio devices used in research studies.
  • Founding Software Engineer, ML, Hearable2025 – 2026Got Transformer audio models running on iPhone with Core ML and INT8 quantisation.

    Hearable. Cambridge, UK. Dec 2025 – Apr 2026.

    • Converted Transformer-based audio models from PyTorch to Core ML for on-device inference on iPhone, running on Apple silicon.
    • Quantised models from FP32 to INT8 to cut model size and inference latency, and verified on-device output parity against the PyTorch originals.
    • Designed training pipelines for audio and signal-based models on noisy real-world data, covering preprocessing, augmentation, batching and validation.
    • Developed FastAPI services to expose models through APIs.
  • Founding Full Stack Engineer, Siddons Yeates and Sharples2025 – 2026Built a 0→1 data and ML platform with FastAPI, PostgreSQL, React and TypeScript.

    Siddons Yeates and Sharples. Contract, Nottingham, UK. Sep 2025 – Jan 2026.

    • Built a 0→1 platform combining data processing and ML-driven workflows with Python, React and TypeScript.
    • Designed backend services with FastAPI and PostgreSQL for data-intensive applications.
    • Developed ranking and scoring systems with statistical and ML methods, evaluated with precision@k and recall@k in iterative feedback loops.
  • Software and Data Engineer, Anteam AI2025Backend systems, data pipelines and embeddings-based retrieval.

    Anteam AI. Internship, Nottingham, UK. Jun 2025 – Jul 2025.

    • Built backend systems and data pipelines for ML-driven analytics, including embeddings-based retrieval and similarity search.
  • Machine Learning Engineer, R&D, The Reciprocal Solutions2023 – 2024Computer vision on NVIDIA Jetson and Coral Edge TPU, fed by real-time Kafka streams.

    The Reciprocal Solutions. Internship, Remote. Dec 2023 – Aug 2024.

    • Deployed computer vision models on NVIDIA Jetson and Google Coral Edge TPU, including INT8 quantisation for the Edge TPU.
    • Built real-time Kafka pipelines streaming CCTV camera feeds into detection and behaviour-analysis models, with a focus on latency and reliability.
  • Software Engineer, Techsnapie Solutions Enterprise2021 – 2022Backend APIs and ETL in Python and SQL.

    Techsnapie Solutions Enterprise. Internship, Remote. Mar 2021 – Sep 2022.

    • Developed backend APIs, data processing and ETL workflows in Python and SQL.

Education

  • MSc Computer Science (Artificial Intelligence), University of Nottingham2025
  • BEng Computer Science and Engineering, Anna University2024

Tools

Languages
Python, TypeScript, JavaScript, SQL and Bash
Frontend
React, Next.js, TypeScript, Tailwind CSS and Redux
Backend
FastAPI, Node.js, Express, Django, REST APIs, GraphQL, WebSockets and model serving
Databases
PostgreSQL, Redis, MongoDB, pgvector, Pinecone, Chroma and FAISS
ML and LLMs
PyTorch, TensorFlow, Hugging Face Transformers, LangChain, LlamaIndex, RAG, LLM APIs, embeddings and similarity search, MLflow and computer vision
On-device ML
Core ML, Apple silicon, ONNX, INT8 quantisation, NVIDIA Jetson and Google Coral Edge TPU
Audio and DSP
audio signal processing, DSP algorithms, audio validation and data ingestion pipelines
Data and streaming
Kafka, ETL pipelines, pandas, NumPy and Airflow
Infrastructure
Docker, Docker Compose, Kubernetes, Terraform, GitHub Actions, CI/CD, AWS and GCP