Data Scientist - Hybrid - Immediate Hiring - BI-ACDAA

Bounteous; Gurugram District, Bengaluru; Hybrid; INR 2,000,000 - 3,500,000; Full time

Job summary

Bounteous is seeking a Data Scientist in India to advance ML-powered solutions with 5+ years of experience. The role focuses on LLM fine-tuning, RAG pipelines, and multimodal AI leveraging Python, PyTorch/TensorFlow, and cloud infra.

You will work across vision-language models, CV tasks, and production deployments with strong collaboration across teams. The ideal candidate will lead architecture decisions, own deployment pipelines, and communicate complex AI concepts to non-technical

Qualifications

• 5+ years hands-on data science / ML / AI engineering, 2–3 years in IC.
• Deep expertise in LLM fine-tuning, LoRA/QLoRA/PEFT, RAG pipelines, embeddings.
• Strong hands-on with Vision-Language Models (LLaVA, GPT-4V, Gemini).
• Proven CV experience: CNNs, transformers, DETR, YOLO, OCR, video understanding.
• Classical ML: ensemble methods, gradient boosting, Bayesian methods, time-series.
• Strong Python and MLDL frameworks: PyTorch, TensorFlow, HF Transformers, LangChain.
• Production ML: deploying, monitoring with SLA, CI/CD for ML, containers.
• Cloud platforms AWS/GCP/Azure; GPU-accelerated training infra.
• Lead cross-functional teams; drive architecture decisions; timely delivery.
• Excellent communication; translate AI concepts for diverse audiences.

Skills

Data science

Machine learning

AI engineering

Python

Communication skills

Leadership

Multimodal AI

LLMs

CNN/Transformers

Production deployment

Education

Master’s or Ph.D. in CS/ML

Tools

PyTorch

TensorFlow

Hugging Face

LangChain

Docker

Kubernetes

AWS

GCP

Azure

Pinecone

Weaviate

Milvus

Job description

Hello Everyone,

Greetings from Bounteous!!

We have an excellent Opening for the below Skills:

Job Title: Data Scientist

Experience: 5+ Years

Primary Skills: LLM (Fine-tuning/RAG) + Computer Vision (CNN/Transformers/Object Detection) + Python with PyTorch/TensorFlow + Production ML on Cloud (AWS/GCP/Azure)

Job Location: Bangalore

Job Description

Required Qualifications

• 5+ years of hands-on experience in data science, machine learning, or AI engineering, with at least 2 to 3 years in IC capacity.
• Deep expertise in LLMs fine-tuning (LoRA, QLoRA, PEFT), prompt engineering, RAG pipelines, embedding models, and LLM evaluation frameworks.
• Strong hands-on experience with Vision-Language Models (VLMs) such as LLaVA, GPT-4V, Gemini, or similar multimodal architectures.
• Proven track record in computer vision CNNs, transformers (ViT, DETR, SAM), object detection (YOLO, Faster R-CNN), segmentation, OCR, and video understanding.
• Solid command of classical ML techniques ensemble methods, gradient boosting (XGBoost, LightGBM), Bayesian methods, and time-series modeling.
• Strong proficiency in Python and ML/DL frameworks (PyTorch, TensorFlow, Hugging Face Transformers, LangChain, or equivalent).
• Production ML experience building, deploying, and monitoring models in real-world systems with SLA requirements.
• Solid understanding of ML system architecture — feature engineering pipelines, model registries, CI/CD for ML, containerized deployments (Docker, Kubernetes).
• Experience with cloud platforms (AWS, GCP, or Azure) and GPU-accelerated training infrastructure.
• Proven ability to lead cross-functional technical teams, drive architecture decisions, and deliver projects on time.
• Excellent communication skills with the ability to translate complex AI concepts for diverse audiences.
• Master’s or Ph.D. in Computer Science, Machine Learning, Statistics, or a related quantitative field preferred.

Technical Depth Expected

• LLMs & NLP: Tokenization, attention mechanisms, transformer architectures, RLHF/DPO, vector databases (Pinecone, Weaviate, Milvus), chunking strategies, and retrieval-augmented generation.
• VLMs & Multimodal AI: Image-text alignment, contrastive learning (CLIP), multimodal embeddings, visual question answering, and document understanding models.
• Computer Vision: Image classification, object detection, instance/semantic segmentation, pose estimation, depth estimation, generative models (diffusion, GANs), and edge deployment (ONNX, TensorRT).
• Classical ML: Feature engineering, hyperparameter tuning, model selection, cross-validation, drift detection, and explainability (SHAP, LIME).
• Production Systems: Model optimization (quantization, pruning, distillation), A/B testing frameworks, shadow deployments, autoscaling inference endpoints, and latency profiling.