Senior AI Research Scientist

Experience : 5 - 15 years

Bhubaneswar, Delhi - NCR, Remote Working

About the Job

Featured

We are seeking a Senior AI Research Scientist with deep expertise in modern foundation models and advanced neural-network architectures. The ideal candidate understands large language models, diffusion and flow-based generative models, state-space models (SSMs), Transformer-SSM hybrids, mixture-of-experts systems, synthetic-data training, and distributed neural-network inference.

This is a hands-on research role for someone who can move comfortably between mathematical theory, rigorous experimentation, prototype implementation, and production collaboration. You will investigate new architectures and training methods while helping build systems that operate efficiently across GPUs, CPUs, NPUs, personal computers, edge devices, private servers, and distributed clusters.

Your work should lead to measurable improvements in model quality, reasoning, latency, throughput, memory use, energy efficiency, privacy, reliability, and total cost of operation. You will have significant influence over product's long-term technical direction and research roadmap.

Key Research Areas

The role will contribute across several of the following areas, with deep specialization expected in at least two:

- Large language models and multimodal foundation models

- Transformer alternatives and attention-efficient architectures

- State-space models, including selective SSMs and Mamba-style architectures

- Hybrid Transformer-SSM, recurrent, sparse, and modular model designs

- Diffusion, discrete diffusion, flow matching, and multimodal generative systems

- Mixture-of-experts models, expert routing, modular networks, and conditional computation

- Synthetic data, model-generated supervision, self-training, and knowledge distillation

- Distributed inference across heterogeneous and intermittently available devices

- Memory-efficient inference, KV-cache management, long-context execution, and model sharding

- Quantization, sparsity, pruning, low-rank adaptation, and dynamic adapter routing

- Agentic models, tool use, planning, reasoning, and multi-agent coordination

- Continual learning, personalization, privacy-preserving learning, and edge AI

Core Responsibilities

Frontier Model Architecture Research

- Design, implement, and evaluate new neural-network architectures for language, reasoning, multimodal generation, and agentic workloads.

- Research the strengths and limitations of Transformers, SSMs, diffusion models, recurrent architectures, mixture-of-experts models, and hybrid designs.

- Develop architectures that combine attention, state-space mechanisms, recurrence, memory, sparse routing, retrieval, and modular components where appropriate.

- Explore effective long-context methods, external and recurrent memory, adaptive computation, speculative execution, and improved reasoning techniques.

- Investigate diffusion and flow-based approaches for text, image, audio, video, structured data, and multimodal generation.

- Translate promising research papers and mathematical concepts into reproducible prototypes and production-relevant experiments.

Synthetic Data and Model-Generated Training

- Design scalable pipelines for creating high-quality synthetic examples, reasoning traces, preferences, critiques, simulations, and task-specific training data.

- Develop teacher-student, self-training, rejection-sampling,



curriculum-learning, process-supervision, and knowledge-distillation approaches.

- Evaluate alignment and post-training methods such as supervised fine-tuning, preference optimization, reinforcement learning, and AI-generated feedback.

- Build filtering, scoring, deduplication, provenance, contamination-detection, and quality-control systems for synthetic datasets.

- Study and reduce the risks of feedback loops, bias amplification, hallucinations, overfitting, reward hacking, and model collapse caused by poorly controlled synthetic data.

- Establish methods for combining synthetic, public, licensed, customer-authorized, and human-generated data while maintaining privacy and traceability.

Distributed Inference and Neural Systems

- Develop algorithms that partition, route, and execute neural-network workloads across heterogeneous devices and infrastructure.

- Research tensor, pipeline, expert, sequence, and context parallelism for inference in resource-constrained and geographically distributed environments.

- Design efficient methods for model sharding, layer placement, distributed KV-cache management, cache-aware scheduling, and dynamic workload migration.

- Improve inference across mixed hardware, including data-center GPUs, consumer GPUs, CPUs, NPUs, Apple Silicon, integrated graphics, edge devices, and browser-based runtimes where appropriate.

- Develop routing and scheduling strategies that account for memory capacity, bandwidth, latency, thermal limits, energy use, device availability, privacy rules, and workload priority.

- Create resilient inference methods that tolerate node loss, unreliable connectivity, changing resource availability, and partial system failure.

- Explore decentralized or collaborative neural networks in which multiple devices jointly execute models without requiring all data or model components to reside in one location.

- Advance compression and acceleration methods, including quantization, sparsity, pruning, distillation, speculative decoding, optimized kernels, and hardware-aware model design.

Research Evaluation and Scientific Rigor

- Form clear hypotheses, define baselines, design ablation studies, and run statistically sound experiments.

- Create evaluation suites covering accuracy, reasoning, robustness, safety, privacy, latency, throughput, memory, energy use, resilience, and cost.

- Identify where conventional benchmarks fail to predict real-world performance and develop task-relevant evaluations.

- Analyze quality-versus-efficiency tradeoffs and produce evidence that guides architecture and product decisions.

- Maintain reproducible research code, experiment records, model cards, dataset documentation, and technical reports.

- Monitor relevant research and clearly communicate which developments are promising, immature, or unsuitable for systems.

Technical Leadership and Product Collaboration

- Help define the company's research roadmap, technical strategy, and standards for scientific quality.





- Collaborate with engineering teams to move successful research from prototype to reliable production systems.

- Work with product leaders to connect research goals to customer needs, deployment constraints, and measurable business or social outcomes.

- Mentor researchers and engineers, review experimental designs, and raise the technical quality of the broader team.

- Contribute to patents, peer-reviewed publications, open research, technical demonstrations, grant proposals, and strategic partnerships when appropriate.

- Explain complex research clearly to technical teams, customers, partners, investors, and nontechnical stakeholders.

Required Skills

- Experience with distributed training or inference systems and parallel-computing frameworks.

- Hands-on work with model parallelism, expert parallelism, distributed KV caches, inference schedulers, or heterogeneous clusters.

- Experience with architectures such as Transformers, selective SSMs, Mamba-style systems, mixture-of-experts models, diffusion models, or hybrid combinations.

- Knowledge of inference engines and optimization stacks such as vLLM, SGLang, TensorRT-LLM, DeepSpeed, Ray, Triton, CUDA, ROCm, MLX, ONNX Runtime, WebGPU, or similar technologies.

- Experience with quantization, low-rank adaptation, dynamic adapters, pruning, sparsity, speculative decoding, or custom kernels.

- Experience creating or governing synthetic datasets at scale, including quality scoring, safety filtering, provenance, and contamination controls.

- Familiarity with alignment and post-training techniques, including preference optimization, reinforcement learning, AI feedback, red teaming, and safety evaluation.

- Experience with multimodal models spanning language, images, audio, video, documents, or structured data.

- Knowledge of privacy-preserving, federated, decentralized, on-device, or edge machine learning.

- A record of leading ambiguous research projects and helping others turn exploratory work into reliable systems.

Qualifications

- PhD in computer science, machine learning, artificial intelligence, applied mathematics, computational science, electrical engineering, physics, or a closely related field; or equivalent research and industry experience.

- Typically 5+ years of relevant research or advanced development experience, including substantial work with modern foundation models or generative AI.

- Deep expertise in at least two of the following: LLM architecture, SSMs, hybrid neural architectures, diffusion or flow-based models, synthetic-data training, model optimization, or distributed inference.

- Strong understanding of the mathematics underlying deep learning, including optimization, probability, linear algebra, numerical methods, sequence modeling, and representation learning.

- Demonstrated ability to convert research ideas into working implementations and evaluate them against strong baselines.

- Advanced proficiency with Python and at least one major machine-learning framework, such as PyTorch or JAX.

- Experience training, fine-tuning, evaluating, or serving large neural networks.

- Understanding of memory, compute, bandwidth, latency, and numerical-precision constraints in modern AI systems.

- Strong experimental judgment, technical writing, and communication skills.

- Evidence of meaningful research impact through publications, patents, open-source contributions, deployed systems, or equivalent technical achievements.

📌 Senior AI Research Scientist (Bhubaneswar)
🏢 Mindfire Solutions
📍 Bhubaneswar