GenAI Platform​/LLM Inference Optimization Engineer; Cloud

Overview

Infosys Data and Analytics (DNA) unit is at the forefront of transforming data into actionable insights, driving business growth and operational efficiency. We specialize in leveraging advanced AI and analytics to create innovative solutions that address complex business challenges. The team pioneers data‑driven decision‑making, enabling organizations to unlock new opportunities and achieve sustainable success. Join a dynamic team that is revolutionizing how businesses harness the power of data and AI.

At Infosys DNA, you’ll work with cutting‑edge technologies, collaborate with industry experts, and contribute to transformative projects that shape the future of business. We foster a culture of continuous learning and growth, ensuring team members thrive in a dynamic, supportive environment.

• Develop data preparation tasks, identifying patterns or anomalies.
• Ensure data readiness for advanced modeling.
• Develop models for complex use cases, refine algorithms to meet business needs, and deploy scalable, production‑ready solutions.
• Conduct testing, optimize algorithms for performance, reliability, and scalability, and guide teammates in best practices.
• Design and develop predictive models and data‑driven analyses to address business challenges.
• Build, evaluate, and deploy models, standardize code, and contribute to knowledge management.
• Leverage tools like SAS and R/Python to create reusable customizations for non‑ML, ML, and deep learning algorithms, enhance analytics including LLMs, and create innovative, cost‑effective solutions.
• Define analytics problems, execute visualization, analysis, and predictive modeling under guidance.
• Proactively maintain models, implement improvements for accuracy and reliability.
• Apply governance controls to mitigate risks and ensure compliance.
• Analyze performance trends, recommend improvements, and document discrepancies for escalation.
• Maintain comprehensive documentation standards and participate in knowledge transfer sessions.
• Participate in discussions with stakeholders to refine requirements, provide insights, and guide implementation of models.
• Apply the predefined quality measurement framework at an individual task level in the project.
• Deploy complex analytics tools or multi‑system integration and validate deployment success.
• Develop scripts or templates for repeated deployment tasks.
• Contribute to analytic solutions, IP asset creation, and training initiatives.
• Contribute to thought leadership such as papers, innovative non‑ML, ML, deep learning or LLM models, and proofs of concepts.
• Deliver analytics training and contribute to content creation.
• Provide input for segment and unit‑level business plans.

• Deliver scalable, high‑quality analytics solutions aligned to business needs.
• Optimise deployment and performance of models.
• Drive innovation through advanced analytics, automation, and thought leadership.
• Enable team growth through knowledge sharing, training, and standardization.
• Support business planning with data‑driven insights.

• vLLM, TensorRT‑LLM, Triton, SGLang.
• Quantization (FP8/AWQ/GPTQ), tensor parallelism.
• Performance benchmarking & tuning.
• Kubernetes, GKE, KServe / ML serving patterns.
• Helm, Operators.
• GPU orchestration concepts and scheduling patterns.
• GCP and/or Azure (strong hands‑on).
• Terraform.
• Cloud networking, landing zones, governance/org policies.
• Hashi Corp Vault (secrets management).

• Prometheus/Grafana, logging, tracing.
• SRE/SLO mindset, reliability engineering.

• Experience in Big Data technologies (e.g., Big Query, Hadoop).
• Expertise in ML model development, data engineering, and software engineering principles.
• Knowledge of MLOps and AI/ML deployment (e.g., Sage Maker, Snowflake); familiarity with CI/CD, Dev Ops, and automation tools in AI/ML contexts.
• Design and implement LLM inference serving stacks using vLLM, TensorRT‑LLM, Triton Inference Server, SGLang.
• Inference optimization techniques: continuous batching, speculative decoding, KV/prefix caching.
• Quantization: FP8 / AWQ / GPTQ and tuning for GPU utilization.
• Build…