Krutrim

As a Developer/Senior Developer at Krutrim, you will be responsible for developing embedded software for different processors. You should have good knowledge in developing drivers for various hardware blocks such as PCIe, I2C, USB, UART, Ethernet, and Crypto security. Understanding Linux kernel internals and working with open-source software is essential for this role. You will be involved in developing high and low-level designs, drivers, and firmware for different hardware blocks. Additionally, you will work on adopting operating systems and embedded software for various processor architectures. Your responsibilities will include developing software based on pre-silicon development vehicles, bringing up software using pre-silicon vehicles and silicon-based platforms, upstreaming open-source code, developing software component-level tests, and enabling them in a CI/CD system. You will also be debugging issues using standard hardware/software-based debuggers and diagnostic equipment. To excel in this role, you should possess excellent knowledge of Linux internals, different drivers, and standards like PCIe, Ethernet, and CXL. Expertise in the Software Development Life Cycle (SDLC), firmware BSP, device drivers, and strong technical problem-solving skills in areas like system boot, UEFI, and OS functionality are crucial. You should have software development skills in C, C++, and Python, along with strong low-level debugging capabilities. A successful candidate for this position should hold a BTech/MTech in Computers, Electronics, or Electrical Engineering and have around 5-12 years of experience in embedded software development across different architectures. Having a good understanding of different CPU architectures like IA, ARM, and RISC V, as well as the Pre-Silicon Development environment, will be advantageous in fulfilling the requirements of this role.,

You are an experienced Lead Generative AI Engineer responsible for training, optimizing, scaling, and deploying various generative AI models including large language models, voice/speech foundation models, vision, and multi-modal foundation models using cutting-edge techniques and frameworks. Your role involves architecting and implementing state-of-the-art neural architecture, robust training, and inference infrastructure to efficiently take complex models with billions of parameters to production while optimizing for low latency, high throughput, and cost efficiency. Your key responsibilities include: - Architecting and refining foundation model infrastructure to support optimized AI models deployment focusing on C/C++, CUDA, and kernel-level programming enhancements. - Implementing optimization techniques like quantization, distillation, sparsity, streaming, and caching for model performance enhancements. - Spearheading the development of Vision pipelines to ensure scalable training and inference workflows of billions of parameter foundation models. - Innovating for state-of-the-art architectures involving Panoptic Segmentation, Image Classification, and Image Generation. - Designing, developing, and innovating state-of-the-art large multimodal models. - Executing training and inference processes to minimize latency and maximize throughput utilizing GPU clusters and custom hardware. - Integrating and tailoring frameworks like PyTorch, TensorFlow, DeepSpeed, Lightening, FSDP, and Habana for fast model training and inference. - Enhancing post-deployment mechanisms with exhaustive testing, real-time monitoring, and robustness checks. - Driving continuous improvement initiatives for deployed models with automated pipelines for drift detection and performance degradation. - Leading the charge in model management including version control, reproducibility, and lineage tracking. - Cultivating a culture of high-performance computing and optimization within the AI/ML domain. Qualifications: - Ph.D. with 5+ years or MS with 8+ years of experience in ML Engineering, Data Science, or related fields. - Demonstrated expertise in high-performance computing with proficiency in Python, C/C++, CUDA, and kernel-level programming for AI applications. - Extensive experience in optimizing training and inference for large-scale AI models. - Understanding of Diffusion Models, Variational Autoencoders, Bayesian Modelling, and Reinforcement Learning is beneficial. - Experience in building billions of parameters generative AI foundation models. - Proven success in deploying optimized ML systems on a large scale using cloud infrastructures and GPU resources. - In-depth understanding and hands-on experience with advanced model optimization frameworks and MLOps tools. - Familiarity with contemporary MLOps frameworks and their application in production environments. - Strong grasp of state-of-the-art ML infrastructures, deployment strategies, and optimization methodologies. - Innovative problem-solving skills and collaborative mindset. - Exceptional communication and team collaboration skills.,

You will be responsible for leading the verification efforts for complex System on Chips (SoCs) and multi-chiplet System on Packages at Krutrim. This includes developing verification plans, testbench environments, and reusable verification IP. You will work on module, sub-systems, chiplets, and complex multi-chiplet testbench environments. Additionally, you will be involved in developing UVM and C/Python based testbench environments for functional, netlist, and power simulations. You will also contribute to developing reference models using System Verilog/SystemC/Python and improving verification flow and methodology. Debugging design units using simulation and formal tools and methodologies will be part of your responsibilities. Collaboration with Architects, Systems, and Software engineering teams on post-silicon debug in the lab is also expected. To qualify for this role, you should have a Bachelor's/Masters degree in Electronics Engineering and a minimum of 10 years of experience in verification of complex SoC designs. Proficiency in System Verilog, SVA, and UVM for code development is essential. Deep knowledge and experience of metric-driven verification methodology and flow are required. You should have extensive experience with various state-of-the-art simulation and formal tools, as well as expertise in verification of CPU/DPU/GPU based sub-systems. Experience with industry standard protocols such as CXL, PCIe, UCIe, and Ethernet protocols will be beneficial. Join Krutrim to be part of the team that is shaping the future of AI computing through cutting-edge silicon design.,

As a Linguist-AI Training & Evaluation at Krutrim (Ola), you will be responsible for curating high-quality datasets, evaluating language models, and fine-tuning performance using Reinforcement Learning from Human Feedback (RLHF). Your role will involve working on evaluating model outputs, DPO Data Set creation, performing comparative analysis, and benchmarking Krutrim's AI models against industry standards. The ideal candidate should possess native-level fluency in Kannada language with excellent command over English. Proficiency in multilingual linguistics is considered a plus. Additionally, a basic understanding of Natural Language Processing (NLP), machine learning models, and large language models (LLMs) is required. You should have proven experience in evaluating AI models, including benchmarking, fine-tuning, and model assessment. An attention to detail is crucial in identifying linguistic nuances, model errors, and biases in AI model outputs. An analytical mindset is also essential for performing quantitative and qualitative assessments of AI model performance, including the development of evaluation metrics. Being web savvy and tech-savvy is advantageous, as you will need good computer skills for managing large-scale data, Kannada typing, and using Microsoft Office suite, among other tasks. The qualifications required for this role include a minimum of a Bachelor's or Master's degree in Journalism, Literature, Linguistics, or a related field from a reputed Indian university. Additionally, 1 to 3 years of experience in translation, localization, content writing, NLP, or AI model evaluation is preferred. This is a full-time role based in Bangalore, with work from the office. Immediate joiners are preferred for this position at Krutrim (Ola).,

As a Principal Research Scientist focusing on AI Alignment at Ola Krutrim in Bangalore, India, you will lead the efforts in Trust and Safety, Interpretability, and Red Teaming within the AI division. Your role will be crucial in ensuring that the AI systems developed are safe, ethical, interpretable, and reliable, with a significant impact on millions of lives. You will be at the forefront of cutting-edge AI research, guiding the implementation of technologies that adhere to the highest standards of safety and transparency. Your responsibilities will include providing strategic leadership for the AI Alignment division, overseeing teams dedicated to Trust and Safety, Interpretability, and Red Teaming. You will work closely with the Lead AI Trust and Safety Research Scientist and Lead AI Interpretability Research Scientist to align goals and methodologies. Developing comprehensive strategies for AI alignment, integrating advanced safety and interpretability techniques, and establishing best practices for red teaming exercises to identify vulnerabilities will be key aspects of your role. Moreover, you will collaborate with product and research teams to implement safety and interpretability aspects throughout the AI development lifecycle. Staying updated on AI ethics, safety, and interpretability research, representing the company in industry events, and managing resource allocation and strategic planning for the AI Alignment division are also part of your responsibilities. Mentoring and developing team members, fostering innovation, and communicating progress and recommendations to executive leadership will be essential in this role. To qualify for this position, you should hold a Ph.D. in Computer Science, Machine Learning, or a related field with a focus on AI safety, ethics, and interpretability. With at least 7 years of experience in AI research and development, including 3 years in a leadership role, you should have expertise in AI safety, interpretability, and red teaming methodologies. Strong knowledge of advanced techniques such as Reinforcement Learning, Proximal Policy Optimization, and attention-based methods, along with experience in overseeing red teaming exercises for AI systems, are required. Your visionary mindset, along with excellent communication skills, project management abilities, and a proven track record in AI safety, ethics, and interpretability research, will be instrumental in shaping the future of responsible AI development at Ola Krutrim. By leading cross-functional initiatives and fostering a culture of continuous learning and innovation, you will contribute to building public trust in AI technologies and positioning the company as a leader in ethical and responsible AI development.,