Artificial Intelligence Research Scientist

Join Us!
Amity, through its AI Research and Application Center, is advancing the frontier of applied AI research across natural language processing, large language models, agentic systems and generative AI. We build intelligent products that serve millions of users and partner organisations across Southeast Asia and Europe. Our research engineers sit at the intersection of scientific inquiry and production engineering—discovering new methods, validating them rigorously, and shipping them into real-world products.

Your Impact:
As an AI Research Engineer at Amity AI Research and Application Center you will own ambitious research goals while ensuring that breakthroughs translate into scalable, production-grade systems.

You will:
Identify high-impact research problems, formulate hypotheses, design experiments and advance the state of the art in areas aligned with the lab’s mission.
Publish findings at top-tier conferences and top-tier leaderboard and contribute to the broader AI research community.
Bridge the gap between research prototypes and production systems, ensuring novel methods are robust, efficient and deployable at scale.
Shape the lab’s research roadmap and propose initiatives that create measurable business and societal value.
Mentor junior researchers and engineers, fostering a culture of scientific rigour and collaborative innovation.

Your Day-to-Day:
Research & Experimentation
Conduct original research in one or more areas: large language models, NLP, computer vision, reinforcement learning, generative models, agentic AI or multimodal learning.
Design and run rigorous experiments—including ablation studies, benchmark evaluations and statistical analyses—to validate new methods and architectures.
Survey, reproduce and extend state-of-the-art results from recent literature; maintain a reading group culture within the team.
Develop novel algorithms, model architectures and training strategies that push performance boundaries on real-world tasks.

Model Development & Optimisation
Design, train and fine-tune large-scale deep learning models (LLMs, diffusion models, multi-modal models) using modern frameworks such as PyTorch, TRL, Unsloth or verl. (Reinforcement Learning Experience is plus)
Optimise model performance through techniques such as knowledge distillation, quantisation, pruning, mixed-precision training and efficient attention mechanisms.
Build and improve training infrastructure for distributed, large-scale model training across GPU/TPU clusters.
Develop evaluation frameworks and metrics to systematically measure model quality, safety and robustness.

Applied Research & Productionisation
Translate research outcomes into production-ready features—building proof-of-concepts (PoCs), prototypes and scalable AI services.
Design and operate RAG pipelines (ingestion, chunking, embeddings, hybrid search, re-rankers) with vector databases (pgvector, Pinecone, Weaviate, OpenSearch) to support retrieval-augmented applications.
Architect and ship LLM-powered agents and chatbots using agentic patterns (tool/function calling, planning, memory, multi-agent orchestration) with robust safety and fallback mechanisms.
Collaborate with product and engineering teams to integrate AI capabilities into customer-facing platforms via APIs and microservices.

Data & Infrastructure
Curate, clean and build high-quality datasets for pre-training, fine-tuning and evaluation; design data pipelines for continuous data collection and annotation.
Implement and maintain scalable ML infrastructure using Docker, Kubernetes, CI/CD and experiment-tracking tools (MLflow, Weights & Biases, or similar).
Monitor deployed models, design automated retraining pipelines and ensure ongoing model quality through observability and alerting.

Knowledge Sharing & Community
Author technical papers, internal reports and blog posts that communicate research findings to both technical and non-technical audiences.
Present research at internal seminars, external conferences and community meetups.
Contribute to open-source projects and public benchmarks to enhance Amity’s visibility in the research community.

Your Ideal Profile:
Required
Education: Master’s or Ph.D. in Computer Science, Machine Learning, Statistics, Mathematics, Electrical Engineering or a related quantitative field.
Research Experience: 3+ years of hands-on experience in AI/ML research or research engineering, with demonstrated ability to design experiments, analyse results and iterate on methods.
Publication Track Record: At least one first-author or co-author publication at a recognised venue (NeurIPS, ICML, ICLR, ACL, EMNLP, CVPR, AAAI, or equivalent), or equivalent demonstrated research output (patents, technical reports, significant open-source contributions).
Deep Learning Expertise: Strong proficiency with Python and modern deep learning frameworks (PyTorch, JAX, or TensorFlow); solid understanding of model architectures (Transformers, diffusion models, GNNs) and training techniques (RLHF, DPO, SFT, pre-training).
Engineering Rigour: Ability to write clean, maintainable, production-quality code; familiarity with software engineering best practices (version control, code review, testing, CI/CD).
Mathematical Foundations: Strong grounding in linear algebra, probability, statistics, optimisation and information theory.

Preferred:
Agentic & LLM Systems: Experience designing agentic architectures (tool/function calling, planning, memory, multi-agent orchestration via frameworks such as LangChain, LlamaIndex, AutoGen or CrewAI).
RAG & Knowledge Systems: Hands-on experience with retrieval-augmented generation pipelines, embedding models, hybrid search and vector databases.
Distributed Training: Experience with large-scale distributed training across multi-GPU/TPU environments (DeepSpeed, FSDP, Megatron-LM or similar).
Cloud & MLOps: Working knowledge of cloud platforms (AWS, GCP or Azure) and ML operations tooling (MLflow, W&B, Kubeflow).
Open-Source Contr