Clari
About this role
The senior agentic AI seat at Clari and Salesloft, and the posting is emphatic that it wants a hands on builder of production agents rather than a researcher experimenting with frameworks on the side. The scope is the machinery that turns an LLM call into a reliable autonomous system: the reasoning and execution loop, the harness managing tool calls, retries, timeouts and session state, short and long term memory, and a skill and tool registry using MCP style tool calling. Beyond that sits planning and multi step reasoning, guardrails and evaluation, and classical statistical and time series work on revenue forecasting and risk models. Eight years in data science or machine learning, of which at least two to three must be building and operating production agents. Bengaluru, hybrid.
Who this is for
What the posting asks for:
- 8+ years in data science or machine learning.
- Within that, at least 2 to 3 years hands on building and operating production AI agents. The posting is explicit that this means production systems, not experimenting with agent frameworks, building demos, or wrapping a single prompt with a tool call.
- Staff or lead level experience is preferred.
Day to day:
- Agent architecture and technical strategy: defining the roadmap for the agentic stack covering the execution loop, harness, memory and tool layers, and deciding when an agent, a classical model or a hybrid is the right answer.
- Production agent engineering: building and operating the reasoning and execution loop, the harness that manages tool calls, retries, timeouts and session state, short and long term memory, and the skill and tool registry using MCP style tool calling.
- Planning and multi step reasoning: designing task decomposition and planning strategies such as evidence based planning, plan and execute, and multi hop reasoning.
- Guardrails, trust and evaluation: owning the evaluation framework for agent behaviour including offline eval, LLM as judge and online A/B testing, plus input and output validation and safety checks.
- Generative AI and revenue modelling: applying statistical and time series methods to forecasting, deal health and risk prediction models.
- Multi agent and cross system coordination: designing how agents communicate with sub agents and other systems.
Location and office reality: the posting states this is a hybrid position in Bengaluru requiring onsite presence as needed, and that candidates must be based in India.
Honest fit guidance: the distinction the posting draws is the one to take seriously. Demos and framework experiments will not clear this bar, and the interview will look for agents you have kept running in production, including how they failed and what you did about it.
- 8+ years in data science or machine learning.
- Within that, at least 2 to 3 years hands on building and operating production AI agents. The posting is explicit that this means production systems, not experimenting with agent frameworks, building demos, or wrapping a single prompt with a tool call.
- Staff or lead level experience is preferred.
Day to day:
- Agent architecture and technical strategy: defining the roadmap for the agentic stack covering the execution loop, harness, memory and tool layers, and deciding when an agent, a classical model or a hybrid is the right answer.
- Production agent engineering: building and operating the reasoning and execution loop, the harness that manages tool calls, retries, timeouts and session state, short and long term memory, and the skill and tool registry using MCP style tool calling.
- Planning and multi step reasoning: designing task decomposition and planning strategies such as evidence based planning, plan and execute, and multi hop reasoning.
- Guardrails, trust and evaluation: owning the evaluation framework for agent behaviour including offline eval, LLM as judge and online A/B testing, plus input and output validation and safety checks.
- Generative AI and revenue modelling: applying statistical and time series methods to forecasting, deal health and risk prediction models.
- Multi agent and cross system coordination: designing how agents communicate with sub agents and other systems.
Location and office reality: the posting states this is a hybrid position in Bengaluru requiring onsite presence as needed, and that candidates must be based in India.
Honest fit guidance: the distinction the posting draws is the one to take seriously. Demos and framework experiments will not clear this bar, and the interview will look for agents you have kept running in production, including how they failed and what you did about it.
Apply on company site
Opens www.salesloft.com, the employer's own application page. Applying is always free.