5 months ago
Toronto, CanadaStaff+
Responsibilities
- Build and operate shared multi-model serving infrastructure across text, voice, code, and vision modalities.
- Own model lifecycle activities including downloading, deploying, serving, monitoring, updating, and swapping models.
- Optimize inference latency, throughput, and cost through quantization, batching, caching, and routing strategies.
- Architect and build runtime infrastructure, orchestration systems, and developer tooling for autonomous AI agents.
- Design multi-agent coordination, safe tool-use infrastructure, workflow automation, permissioning, sandboxing, and human-in-the-loop systems.
- Develop evaluation and observability frameworks, SDKs, and APIs for building and deploying agents.
- Define technical direction and architecture, establish agent design standards, mentor engineers, and lead complex technical initiatives.
Requirements
- 8+ years of software engineering experience, including 3+ years in AI systems or LLM applications.
- Strong understanding of LLM agent architectures, tool use, multi-step workflows, multi-agent coordination, and failure modes.
- Experience building highly reliable distributed systems and evaluating LLM systems in production.
- Proficiency in TypeScript or Python and willingness to work with both; the agent platform uses TypeScript on Bun with Temporal workflows on Kubernetes and EC2, while the inference platform uses Python.
- Experience with modern LLM APIs or open-source models and familiarity with model serving technologies such as vLLM, TensorRT-LLM, or Triton.
- Understanding of task queues, event-driven architectures, state management, and durable long-running workflows.
- Experience with AWS or GCP and containerized deployments.
- Understanding of security risks in agentic systems, including prompt injection, privilege escalation, and data leakage.
- Demonstrated experience leading complex technical initiatives and strong written and verbal communication skills.
- Preferred qualifications include experience in regulated industries, familiarity with Model Context Protocol, model fine-tuning, quantization, LoRA, CI/CD automation, developer tooling, Temporal, Airflow, Prefect, voice or multimodal models, edge inference, and human-in-the-loop oversight systems.
Benefits
- Paytm Labs is committed to diversity, equal opportunity, accessibility, and accommodation during the recruitment process.
- The role partners across ML, product, security, and other internal teams and supports production systems serving enterprise customers.
- Only applicants selected for an interview will be contacted.
Tech Stack
Categories
About Paytm
Paytm builds a digital payments and financial services platform for consumers and merchants in India, spanning the Paytm app for UPI, wallet and bill pay, plus QR/POS and Soundbox devices for in-store acceptance. It monetizes through payment processing fees, merchant subscriptions, and commissions on financial products such as lending and insurance. Founded in 2010 and headquartered in Noida, it is publicly listed in India and serves over 300M users and 20M merchants.