# Shubham Baid

Shubham Baid is a remote worker on Himalayas, Senior AI Architect at BLUE, based in India.

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

```text
Applied AI Engineer at BLUE building production vision-language systems, including a video compression system that cut median bitrate by 40%.
```

| Detail | Value |
| --- | --- |
| Availability | Looking for a job |
| Location | India |
| Authorized to work in | India |
| Salary expectations | 70k+ USD |
| Job categories | Applied AI Engineer, AI Engineer, Machine Learning Engineer |
| Skills | Applied AI, Agent Evaluation, Vision Language Models, VLM Benchmarking, Custom Task Aligned Metrics, Machine Learning, Computer Vision, Multi Agent Systems, Production ML Systems, Edge AI, Model Fine Tuning, Quantization, Knowledge Distillation, RAG, Vector Databases |
| Website | https://shubhambaid.github.io |
| LinkedIn | https://linkedin.com/in/shubhambaid21 |
| GitHub | https://github.com/shubhambaid |

## About

```text
At BLUE, I developed Foreground VMAF, a task-aligned quality metric that became the team's benchmarking standard, and used semantic metrics to guide video compression changes. The system cut median bitrate by 40% while keeping detection precision and recall changes under 1%.

I also took an edge pipeline from RTSP ingest through VLM inference to production across 1,000+ cameras, and cut VLM calls by 53% with packet-size-based frame skipping. Across my work at BLUE, NeoFab, CAFU, Avathon, and Integration Wizards, I've built and shipped AI systems spanning agent workflows, video analytics, and production computer vision.
```

## Experience

### Senior AI Architect (current)

BLUE. Apr 2025 - Present

```text
Built a benchmarking framework for Vision-Language Models and developed Foreground VMAF, a task-aligned quality metric. Led production edge pipeline across 1,000+ cameras, achieving 40% median bitrate reduction with minimal impact on downstream accuracy.
```

Skills: Vision-Language-Models, Benchmarking, Foreground-VMAF, Video-Compression, Edge-AI, RTSP, GStreamer, Nvidia-DeepStream, TensorRT, CUDA, NVIDIA-Jetson, PyTorch, VLM, H264, H.265

### Multi-Agent Systems Engineer

NeoFab. Apr 2025 - Apr 2026

```text
Built multi-agent factory systems with architectural containment, including BOM extraction, document enrichment, and LLM-based machine reranking. Developed order validation with GPT-4 vision and contributed to manufacturing analytics.
```

Skills: Multi-Agent-Systems, LangGraph, Crew-AI, MCP, RAG, GPT-4-Vision, Prompt-Engineering, Manufacturing-Analytics, OEE-Improvements, Takt-Time, Bottleneck-Analysis

### AI Engineer

CAFU. Jan 2025 - Apr 2025

```text
Re-architected ETA prediction pipeline with real-time behavioral signals, reducing delayed fulfillment from 13% to 8% and severe SLA breaches by 42%. Built agentic LLM workflows for marketing content and multi-modal LLM agents for B2B lead acquisition.
```

Skills: Predictive-ETA, Signals, Agentic-AI, Multi-Modal-LLMs, Content-Marketing, B2B-Acquisition, Python, PyTorch

### Senior AI Engineer

Avathon. May 2022 - Jan 2025

```text
Built a gamified active-learning annotation tool and led AI for the VAIA School Safety Suite with real-time video understanding across 100+ cameras. Trained and shipped detection and classification models, integrated VLMs into classical CV pipelines, and cut inference CPU usage by 50%.
```

Skills: Active-Learning, Annotation-Tools, Real-Time-Video-Understanding, Object-Detection, Classification, Vision-Language-Models, ARM-NEON, INT8-Quantization, Inference-Optimization, PyTorch

### Senior AI Engineer

Integration Wizards. Sep 2020 - May 2022

```text
Owned a production ALPR system end to end and migrated legacy CV models to DeepStream and TensorRT GPU pipelines. Delivered PPE, fall-arrester, and vehicle classification from dataset curation through deployment.
```

Skills: ALPR, YOLO, OCR, Nvidia-DeepStream, TensorRT, GPU-Pipelines, PPE-Detection, Fall-Arrester-Detection, Vehicle-Type-Classification, Dataset-Curation, Deployment

## Education

### REVA University

Bachelor of Technology, Computer Science. 2017 - 2021

Activities and societies: Best Outgoing Student; Founder of GDSC REVA.

```text
Pursued a Bachelor of Technology in Computer Science from REVA University, Bengaluru, from 2017 to 2021. Recognized as Best Outgoing Student and founded GDSC REVA.
```

## Tech stack

- GStreamer
- Python
- C++
- PyTorch
- FastAPI
- Linux
- Docker
- CUDA
- Hugging Face
- CrewAI
- MLX
- LangGraph
- MuJoCo

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

## People also viewed

- [Samuel Davis](https://himalayas.app/@samueldavis): AI Lead, E42
- [Ram Krishna Tiwari](https://himalayas.app/@ramkrishnatiwari): Multi Label Movie Classifier, Movie Genre Classification
- [Dhruvin Meshiya](https://himalayas.app/@dhruvinmeshiya): AI & Machine Learning Engineer, Annotatehub Solutions LLP
- [Anupam Tripathi](https://himalayas.app/@anupamtripathi): Founding AI Engineer, Hashteelab
- [Ganesh Subramanya](https://himalayas.app/@ganeshsubramanya): AI Engineer, Novel Office
- [Afham Ahmad](https://himalayas.app/@afhamahmad): Machine Vision Intern, KlarSYS Automations LLP
- [Jaikishan User](https://himalayas.app/@jaikishan2k): AI Engineer Agentic Systems, L&T Energy & Hydrocarbon
- [suyash verma](https://himalayas.app/@suyashverma): AI / ML Engineer, General Motors Project via TCS
- [MANSI GOEL](https://himalayas.app/@mansigoel): Senior Software Engineer – AI/ML & Agentic Systems, NVIDIA
- [Faheem Thanveer](https://himalayas.app/@faheemthanveer): Senior AI Engineer, Blackstraw.ai
