Senior Research Engineer - AI/ML

immunitoAI Private Limited; Bengaluru; Presencial; INR 3.000.000 - 6.000.000; Jornada completa

Descripción de la vacante

immunitoAI Private Limited in Bengaluru is seeking a Senior Research Engineer - AI/ML to join our early-stage team. You will design deep neural networks from first principles, build production-ready code, and work closely with scientists and bioinformaticians to translate biology problems into ML solutions.

This role emphasizes end-to-end ownership, from architecture design to deployment, with a focus on Generative AI, transformers, and efficient inference.

Formación

• Strong foundation in Deep Neural Networks.
• Hands-on, industry experience writing production-grade ML code (Python).
• Prior experience building custom Deep Neural Networks from scratch.
• Solid foundations in linear algebra, probability, optimization.
• Experience with PyTorch or TensorFlow.
• Familiarity with AI agent-based coding tools (e.g., Claude Code, Copilot).
• Ability to work in a cross-domain team and tackle new problems independently.
• Willingness to learn and explore new avenues.

Responsabilidades

• Designing and implementing deep learning algorithms from scratch.
• Writing extensive code for Training, Inference, Data Analysis, and Processing.

Conocimientos

Deep Neural Networks

Python

PyTorch

TensorFlow

ML Systems

Research Experience

Educación

Master's degree in CS/Math/related field

Herramientas

PyTorch

TensorFlow

Descripción del empleo

Locations

• Bangalore

About This Role

We\'re looking for a Senior Research Engineer - AI/ML to join our research and engineering team at an early-stage start-up working at the intersection of deep learning and biology. This is a hands-on engineering role — you\'ll be designing and building deep neural network architectures from scratch, not fine-tuning off-the-shelf models or running MLOps pipelines. You\'ll work closely with scientists, immunologists, biochemists, and bioinformaticians to translate real biological problems into working machine learning systems, taking ideas from problem definition and architecture design all the way through to trained, deployed models. If you enjoy building things from first principles as much as shipping production-quality code, and want to use modern AI-assisted coding tools to move faster on genuinely hard problems, this role is built for you.

Minimum Qualification

• Master\'s degree in Computer Science, Mathematics, or a related field, plus 3+ years of industry experience building custom Deep Neural Networks

(Note: We are not considering candidates straight out of a Master\'s program with no industry experience, or candidates without a Master\'s degree — this is a senior, experienced-hire role.)

Ideal Candidates

• Experienced engineers with a Master\'s degree in Computer Science, Mathematics, or a related field, and a strong industry track record building Deep Neural Networks.

• Senior engineers with a track record of independently owning ML systems end-to-end, from architecture design through deployment.

• Candidates with deep, hands-on Generative AI experience will be given strong preference: having designed, implemented, and trained generative models such as Diffusion, Flow Matching, Score-based, or Consistency models from the ground up, with a solid understanding of the underlying mathematics. (Experience limited to fine-tuning pre-trained models, prompt engineering, or building API-based GenAI applications will not be considered sufficient.)

• Candidates with in-depth knowledge of modern Transformer and LLM architectures (e.g., attention variants, positional encodings such as RoPE, Mixture-of-Experts, normalization and scaling choices), and hands-on experience training them and running efficient inference, including test-time compute scaling (e.g., sampling and search strategies, reasoning-time compute allocation), will be given strong preference.

• This position requires strong coding expertise using PyTorch and comfort working with AI agent-based coding tools (e.g., Claude Code, Copilot, or similar) to accelerate engineering workflows.

• Applicants with hands-on experience in building Diffusion, Graph Neural Networks, or Point Cloud Architectures will be given higher priority.

Responsibilities

• Designing and implementing deep learning algorithms from scratch to solve unique problems

  • Please note: this is a research and engineering position, not a data analytics or MLOps opening

• Writing extensive code from scratch for Training, Inference, Data Analysis, and Processing — leveraging AI agent-based coding tools where they improve speed and quality

• Translating mathematical formulations and research ideas into robust, production-ready code

• Reviewing relevant literature on existing machine learning and deep learning approaches and evaluating applicability to problems at hand

• Understanding data and refining feature sets

• Mentoring junior engineers and setting engineering best practices for the ML team

• Learning and adapting fast in an early-stage start-up

• Taking end-to-end responsibility for design, development, testing, and release of ML systems and products

• Working in a cross-domain team with software engineers, scientists, immunologists, biochemists, and bioinformaticians

Must Have

• Strong foundation in Deep Neural Networks

• Hands-on, industry experience writing production-grade machine learning code in one or more programming languages (preferably Python)

• Prior experience building custom Deep Neural Networks from scratch

• Strong software engineering and machine learning fundamentals, with solid mathematical grounding (linear algebra, probability, optimization) to read research papers and implement them in code

• Prior experience with PyTorch or TensorFlow

• Familiarity with AI agent-based coding tools (e.g., Claude Code, Cursor, Copilot) and using them effectively in an engineering workflow

• Ability to work and communicate in a cross-domain team

• Ability to explore new problem statements independently

• Openness to learning and exploring new avenues

Good to Have

• Prior professional experience in building custom Diffusion models

• Prior professional experience in building custom Graph Neural Network models

• Prior professional experience in building custom Point Cloud Architectures

• Prior professional or research experience in building custom generative models using Flow Matching, Score-based, or Consistency model approaches

• Hands-on experience training Transformer models from scratch (not only fine-tuning), including large-scale or distributed training (e.g., FSDP, DeepSpeed, mixed precision)

• Experience with LLM inference optimization and test-time compute techniques (e.g., KV-cache optimization, quantization, speculative decoding, best-of-N sampling, tree search, verifier-guided decoding)

• Familiarity with efficient attention implementations (e.g., FlashAttention, sparse or linear attention) and long-context modeling

• Experience applying generative or Transformer-based models to scientific or biological data (e.g., protein, molecular, or sequence modeling)

• Experience mentoring or leading small engineering teams

• Published research (peer-reviewed papers, preprints) in relevant deep learning areas