LLM Engineer

About Company:

A Forex Trading Product Client of Adecco based in Hyderabad

An Ideal Candidate:

• Seeking a highly technical LLM / AI Engineer to build, optimize, and deploy the models that power our Agentic AI initiatives. You will be the primary person responsible for ensuring our agents don't just "chat," but execute with precision.
• Your focus will be on model selection, prompt engineering at scale, RAG (Retrieval-Augmented Generation) optimization, and ensuring our AI output is grounded in our specific business logic.
• This role requires a "builder" mentality, someone who can take a raw model and turn it into a production-ready asset that significantly shortens our support cycles. 

Key Competencies:

• Model Engineering: Implement and optimize LLM pipelines using both proprietary (OpenAI, Anthropic) and open-source models (Llama, Mistral) to find the best cost-to-performance ratio.
• RAG Architecture: Build and maintain sophisticated RAG systems to ensure agents have real-time access to our internal documentation, customer history, and trading platform data.
• Prompt Engineering & Tuning: Develop and version-control complex prompt templates, utilizing techniques like Few-Shot prompting and Chain-of-Thought to improve reasoning.
• Evaluation & Testing: Establish "LLM-as-a-judge" frameworks and automated testing to measure hallucination rates, toxicity, and accuracy before agents hit the non-production environments.
• Tool-Use Logic: Program the logic that allows agents to use external tools (APIs, Database queries, Moltbot functions) reliably and safely.
  • -Latency Optimization: Streamline model inference and data retrieval to ensure the "Super Employee" responds in near real-time, improving customer satisfaction. - A background in Backend Engineering or Data Science with a heavy focus on NLP.
  • Architecture: Deep experience in scalable system design and distributed architecture.
  • - Languages: Python (Advanced) and experience with data libraries (NumPy, Pandas).
  • - AI Frameworks: LangChain, LlamaIndex, or Haystack.
  • - Vector Infrastructure: Experience with Vector Databases and semantic search (FAISS, Chroma, or Pinecone).
  • - Embedding Models: Deep understanding of text embeddings and how to optimize them for domain-specific context (Trading/FinTech).