# Devaryan User

Devaryan User is a remote worker on Himalayas, AI Developer (Full-Stack and AI Engineering) at Metis, based in United Kingdom.

Source: https://himalayas.app/@devaryanuser

```text
AI Engineer at Metis who shipped an end-to-end RAG pipeline and built a speech turn-detection benchmark.
```

| Detail | Value |
| --- | --- |
| Availability | Looking for a job |
| Location | United Kingdom |
| Authorized to work in | United Kingdom |
| Salary expectations | 60k+ GBP |
| Job categories | AI Engineer, Machine Learning Engineer, Full Stack Developer |
| Skills | LLM Engineering, RAG, LLM Evaluation, Model Evaluation Harnesses, Data Ingestion, Chunking, Embeddings, Hybrid Retrieval, Vector Similarity Search, Reranking, Prompt Engineering, Token Budgeting, Context Window Optimization, Provider Abstraction, Retry Strategies |
| LinkedIn | https://linkedin.com/in/devaryan-devaryan |
| GitHub | https://github.com/Devaryan-one-1 |

## About

```text
At Metis, I shipped a production RAG pipeline covering ingestion, retrieval, prompt construction, and the OpenAI API layer. I traced poor answer quality to fixed-size chunking that separated financial tables from their headers and footnotes, then rebuilt chunking around document structure.

I also implemented hybrid Elasticsearch and dense-vector retrieval with reranking, and built a provider abstraction with token budgeting, retries, and failure handling. I defined accuracy KPIs and reported where outputs were ready for production and where human review remained necessary.

On my Voice-AI Turn-Detection Benchmark project, I evaluated six strategies across 1,152 labelled speech turns, measuring latency alongside interruption rate. I caught and corrected a latency measurement bug, retracted two claims after paired reanalysis, and built an async testable pipeline.
```

## Experience

### AI Developer (Full-Stack and AI Engineering)

Metis. Sep 2024 - Jan 2025

```text
Shipped a production RAG pipeline end to end, owning ingestion, chunking, embedding, retrieval, prompt construction, and the OpenAI API call layer. Diagnosed poor answer quality as an ingestion failure and rebuilt chunking around document structure, improving retrieval relevance and downstream answer quality. Implemented hybrid retrieval with Elasticsearch and dense vector similarity, added rerank
```

Skills: RAG, Ingestion, Chunking, Embeddings, Retrieval, Prompt-Engineering, OpenAI-API, Elasticsearch, Hybrid-Retrieval, Reranking, Token-Budgeting, Retry-handling, Latency-Optimization, KPI-Tracking

### AI Developer Intern

Metis. Jun 2024 - Sep 2024

```text
Built machine learning models from scratch and integrated them into production web applications, architecting modular backend infrastructure and abstraction layers. Built application layers, REST APIs, and interfaces connecting the full-stack system to model backends, and developed visualisation systems for interpretability.
```

Skills: Python, TypeScript, Next.js, React, Flask, Django, NodeJS, OpenAI-API, Elasticsearch, PyTorch, AWS, Git

## Education

### University of Birmingham

Master of Science, Computer Science. 2025 - Present

```text
Pursuing an MSc in Advanced Computer Science at the University of Birmingham from 2025 to 2026.
```

### Indian Institute of Information Technology, Lucknow

Bachelor of Technology, Computer Science & AI. 2020 - 2024

```text
Completed a BTech in Computer Science & AI at the Indian Institute of Information Technology, Lucknow from 2020 to 2024.
```

## Tech stack

- GitHub Actions
- NumPy
- Pandas
- PostgreSQL
- MongoDB
- Node.js
- Django
- Next.js
- React
- JavaScript
- Python
- C++
- PyTorch
- scikit-learn
- Flask
- FastAPI
- asyncio
- Elasticsearch
- AssemblyAI
- TypeScript
- pytest
- Git
- Docker
- SQL
- Ollama
- OpenAI API
- Retrieval-Augmented Generation (RAG)
- Express
- AWS

To contact Devaryan, message them from this profile with a Himalayas employer account: [Hire on Himalayas](https://himalayas.app/hire).

## People also viewed

- [Aamer Hussain](https://himalayas.app/@aamerhussain2): Full Stack Engineer, MicroDevSolutions
- [Tharun kumar](https://himalayas.app/@tharunkumar5): Frontend & AI Intern, KalV Consultants and Facilitators
- [Terry Migwi](https://himalayas.app/@terrymigwi): AI and Data Scientist, Multiverse
- [KRISHNA CHAITANYA GOUD MAKTHALA](https://himalayas.app/@krishnachaitanyagoud): AI Engineer (Contract), Avsar
- [Haruto Saito](https://himalayas.app/@harutosaito): Senior AI Engineer, V7
- [tehilla obanor](https://himalayas.app/@tehillaobanor): Forward Deployed AI Engineer, Afterframe AI Orchestration Platform
- [Arfa Patekari](https://himalayas.app/@arfapatekari): AI Engineer, LLM Evaluation & Benchmarks, Alignerr (Labelbox) and Mercor
- [Usama Khatab](https://himalayas.app/@usamakhatab): Senior Software Engineer, Cambridge University Press & Assessment
- [Sachin Bhatiya](https://himalayas.app/@sachinbhatiya): AI & Data Science Engineer, Inforens
- [Avishek Roy](https://himalayas.app/@avishekroy): Software Engineer (MLOps), ValueMomentum