About AiSensy
AiSensy is a WhatsApp-based customer engagement and marketing platform helping businesses improve communication and revenue growth through WhatsApp.
More than 210,000 businesses use AiSensy for customer engagement and communication workflows.
Billions of WhatsApp messages are exchanged annually through the platform.
Trusted by leading brands across multiple industries.
Businesses using AiSensy have achieved significant improvements in customer engagement and revenue performance.
About the Role
You will work on the machine learning systems powering AiSensy’s intelligent chat platform serving businesses across India. This role involves strong ownership, architectural decision-making, and close collaboration with senior engineering leadership.
This is a production-focused engineering role.
You will work on scalability, response performance, infrastructure efficiency, and overall system reliability.
Core Responsibilities
Conversational Systems
Own the end-to-end ML pipeline for intelligent chat systems including response quality, workflow routing, validation controls, and answer generation.
Design and optimize hybrid search systems using BM25, ColBERT, dense vector representations, and semantic search databases.
Build response validation checks that identify issues before customer impact.
Work across workflow frameworks for request separation, feature selection, process routing, and consistent structured responses.
Improve multi-layer memory systems spanning short-term memory, summaries, and long-term storage using technologies like Qdrant and Valkey.
Build validation systems for sensitive data filtering, usage guidelines, and policy compliance while maintaining separation between platform rules and customer-specific configurations.
Run detailed testing and performance analysis while connecting model quality with business metrics such as resolution rate, escalation rate, and operational cost.
Behavior Modeling
Dand productionize behavior modeling systems that convert multi-turn customer interactions into vector representations suitable for grouped analysis, personalization, and intelligent recommendations.
Build multi-level user grouping pipelines across platform-wide and organization-specific datasets with incremental updates as new information becomes available.
Collaborate with platform engineering teams on infrastructure covering event storage, behavioral feature computation, semantic storage systems, and low-latency online serving.
Own the production ML lifecycle including training, batch inference, real-time inference, retraining schedules, performance monitoring, and controlled recovery mechanisms.
Platform & Production
Deploy ML systems using Amazon Bedrock, SageMaker, and AWS infrastructure including ECS, ECR, and Kubernetes.
Own response time, infrastructure cost, and quality metrics for large-scale conversational workloads operating across shared infrastructure environments.
Write detailed low-level design documents before implementation and participate in technical reviews with engineering teams.
Contribute to internal benchmarking systems evaluating platform quality against leading conversational support products.
What We’re Looking For
Must-Have Skills
5+ years building production ML systems with at least 2 years focused on conversational systems or generative AI applications.
Strong understanding of retrieval-augmented systems including diagnosing search quality issues, optimizing hybrid search strategies, and identifying when retrieval systems are unnecessary.
Hands-on experience with LangGraph, DSPy, or similar workflow frameworks.
Strong understanding of semantic search databases such as Qdrant, pgvector, or equivalent systems at meaningful scale.
Production experience with cloud ML platforms such as SageMaker, Bedrock, Vertex AI, or Azure ML.
Solid understanding of machine learning fundamentals including vector models, clustering, contrastive learning, and evaluation techniques.
Ability to write low-level design documents that can be reviewed and implemented by senior engineering teams.
Strong Python and FastAPI experience.
Preferred Signals
Ability to identify when generative AI is not the right solution for a problem.
Experience building scalable ML systems for shared customer environments where infrastructure efficiency matters alongside model quality.
Experience working with Indic-language models such as MuRIL or IndicBERT.
Experience with user journey analysis or sequential behavior modeling.
Comfort discussing both infrastructure and ML architecture decisions.
Nice to Have
Experience with WhatsApp Business API or CPaaS platforms.
Contributions to open-source ML tooling.
Published work related to conversational systems, search systems, or behavior modeling.
How We Evaluate
We value strong problem-solving and engineering judgment.
Assignments with execution errors will not be considered.
We look for engineers who can break down problems, choose practical solutions, and clearly justify technical decisions instead of overusing generative AI approaches where simpler systems are more effective.
