Research Scientist
First seen by Alion on Sep 11, 2026.
About the Role:
We are looking for a Research Engineer to build the next generation of AI agents capable of solving complex business problems through reasoning, retrieval, and multi-step decision making.
You'll work on challenging research problems involving LLM agents, reinforcement learning, retrieval systems, evaluation frameworks, and continual learning. This role is ideal for engineers who enjoy experimenting with cutting-edge AI techniques and shipping production-ready research.
If you have experience building intelligent AI agents, training LLMs, designing evaluation pipelines, and improving model behavior through experimentation, we'd love to hear from you.
Key Responsibilities:
- Design and improve agentic AI systems capable of complex multi-step reasoning.
- Build intelligent tool-use strategies, planning mechanisms, and multi-agent workflows.
- Develop advanced context engineering and memory management techniques for AI agents.
- Design automated evaluation frameworks, benchmarks, and regression testing pipelines for LLM applications.
- Train and fine-tune large language models using techniques such as LoRA, QLoRA, RLHF, DPO, GRPO, and SFT.
- Research and implement reinforcement learning approaches for improving agent reasoning and decision-making.
- Build retrieval systems using dense retrieval, hybrid search, reranking, GraphRAG, and knowledge graphs.
- Develop explainable scoring models and prediction systems from structured and unstructured data.
- Collaborate with engineering and research teams to rapidly prototype, evaluate, and deploy AI solutions into production.
Required Skills:
- Strong Python programming skills.
- Hands-on experience with PyTorch and Hugging Face Transformers.
- Experience building AI Agents or Multi-Agent systems.
- Deep understanding of LLM fine-tuning and optimization techniques.
- Experience with RLHF, DPO, PPO, ORPO, GRPO, or similar preference optimization methods.
- Strong knowledge of Information Retrieval, Search, RAG, Hybrid Retrieval, or GraphRAG.
- Experience with Agentic Memory, Context Engineering, or Knowledge Graphs.
- Experience designing LLM evaluation frameworks, benchmarking systems, and automated evaluation pipelines.
- Knowledge of Continual Learning or Lifelong Learning techniques.
- Experience with Knowledge Distillation for language or retrieval models.
- Strong understanding of model behavior analysis and experimentation.
Preferred Skills:
- LangGraph, DSPy, OpenAI SDK, Anthropic SDK.
- vLLM, DeepSpeed, TRL, Unsloth.
- Vector Databases (Pinecone, Weaviate, Milvus, Qdrant).
- Neo4j or Knowledge Graph technologies.
- Ray, Kafka, Docker, Kubernetes.
- FastAPI, AWS or GCP.
- MLflow, Weights & Biases, Arize, Langfuse.
Preferred Qualifications:
- Experience building production-grade AI agents.
- Experience with retrieval models, ranking systems, or enterprise search.
- Publications in top-tier AI/ML conferences are a strong plus.
- MS or PhD in Computer Science, Artificial Intelligence, Machine Learning, Statistics, Mathematics, or a related quantitative discipline.
Why Join?:
- Work on cutting-edge Agentic AI and LLM research.
- Solve challenging real-world AI problems at scale.
- Collaborate with a highly technical engineering and research team.
- High ownership, rapid innovation, and strong growth opportunities.
- Competitive compensation and benefits.
Skills
Artificial Intelligence, Machine Learning, Generative AI, Data Science, IT Research, LLM