Senior AI Engineer - LLM/RAG

Description :

We are looking for a Senior AI Engineer to build the next generation of intelligent trading and advisory systems.

You will work at the intersection of Stockbroking Open Source AI, deploying scalable models on AWS that process millions of transactions and market ticks daily.

You will move beyond simple chatbots to build agentic workflows that assist traders, automate compliance, and predict market anomalies in real-time.

What You Will Build (Key Responsibilities) :

Real-Time Financial Models :

- Develop and deploy low-latency models for stock trend prediction, algorithmic trading signals, and fraud detection using PyTorch/TensorFlow

Generative AI for Finance :

- Fine-tune open-source LLMs (Llama 3, Mistral, Gemma) on proprietary financial datasets to build "Market-Aware" RAG (Retrieval-Augmented Generation) systems for research reports and client advisory.

AWS Cloud Architecture :

- Architect serverless inference pipelines using AWS SageMaker, Lambda, and Fargate. optimize costs by utilizing Spot Instances and AWS Inferentia chips.

Data Engineering :

- Build robust ETL pipelines using AWS Glue and Athena to process high-frequency tick data and structured financial reports.

Compliance & Security :

- Ensure all AI models comply with SEBI regulations regarding data privacy. Implement "Privacy-Preserving ML" techniques to ensure customer data never leaves our secure VPC.

The "Standard" Tech Stack (Must Haves) :

Languages :

- Python (Advanced), SQL, C++ (Bonus for high-frequency trading optimization).

Cloud (AWS) :

- SageMaker, Bedrock, Lambda, S3, ECR (Elastic Container Registry).

AI/ML Frameworks :

- PyTorch, Hugging Face Transformers, LangChain/LlamaIndex, Scikit-learn.

Vector Databases :

- Qdrant, Milvus, or AWS OpenSearch (for RAG applications).

DevOps :

- Docker, Kubernetes (EKS), GitHub Actions for CI/CD.

Experience We Value (The Differentiators) :

Financial Domain Knowledge :

- Understanding of technical indicators (RSI, MACD), options Greeks, or fundamental analysis ratios.

Open Source Contributions :

- A history of contributing to or utilizing open-source AI projects (we value engineers who don't just use APIs but understand the underlying code).

Latency Optimization :

- Experience quantifying models (Quantization, Pruning) to run on CPU/Edge devices to reduce cloud costs.

Visualizing the Role :

To help you understand where this role fits, here is how standard Fintech AI teams structure their workflow :

1. Data Ingestion : Market Feeds - AWS Kinesis.

2. Processing : AWS Glue - Feature Store.

3. Training : SageMaker (using Open Source Models).

4. Inference : AWS Lambda (Real-time) for users.