Overview

Trinka AI (www.trinka.ai) is an AI-powered writing assistant designed for academic and technical writing.

Responsibilities

• Design, develop, and deploy AI/ML models with a focus on Large Language Models (LLMs)
• Implement and optimize high-performance model serving solutions (e.g., vLLM, FastAPI, Triton)
• Build and maintain agent-based systems (Agentic AI, RAG pipelines, multi-agent workflows)
• Develop scalable data pipelines for preprocessing, embedding generation, and retrieval
• Work on vector databases (FAISS, ChromaDB, etc.) for similarity search and RAG use cases
• Optimize inference latency, throughput, and cost for production deployments
• Collaborate with cross-functional teams to integrate AI solutions into products
• Monitor, debug, and improve deployed AI systems for reliability and performance
• Stay updated with the latest advancements in LLMs, multimodal AI, and agent frameworks

Qualifications

• Minimum 1 year of hands-on experience working with Agentic AI systems (multi-agent workflows, autonomous agents, or orchestration frameworks)
• 2–5 years of experience in AI / Data Science / Machine Learning
• Strong proficiency in Python and ML frameworks (PyTorch, TensorFlow, Hugging Face)
• Hands-on experience with Large Language Models (LLMs) and prompt engineering
• Experience with vLLM or similar high-performance inference engines
• Practical knowledge of building agent-based systems (LangChain, CrewAI, custom agents, etc.)
• Experience with Retrieval-Augmented Generation (RAG) architectures
• Familiarity with vector databases (FAISS, ChromaDB, Pinecone, etc.)
• Strong understanding of REST APIs and backend frameworks (FastAPI preferred)
• Experience with Docker, Kubernetes, or cloud deployments (AWS/GCP/Azure)
• Solid understanding of data structures, algorithms, and system design
• Experience with multimodal models (LLaVA, image-to-text, OCR pipelines, etc.)
• Knowledge of distributed systems and GPU optimization techniques
• Experience with streaming inference and real-time AI systems
• Familiarity with model quantization, batching, and memory optimization
• Exposure to open-source LLM ecosystems (OpenRouter, vLLM, TGI, etc.)
• Experience with logging, monitoring, and observability tools
• Knowledge of CI/CD pipelines for ML systems (MLOps)

Education

• Minimum qualification: B.Tech / B.E. in Computer Science, Electrical Engineering, or a related field