Boggu Maheshbabu

As a Machine Learning Engineer at Salesken.ai, I drive the development of AI systems that transform customer conversations into meaningful business impact. My work focuses on building scalable, production-grade solutions powered by LLMs, RAG pipelines, and multi-agent systems, enabling more personalized, context-aware, and efficient interactions for end users.

I have designed and deployed multi-agent architectures using LangGraph, LangChain, and OpenAI models—optimizing how our product retrieves, critiques, and delivers information. These innovations have directly improved customer satisfaction, product adoption, and engagement outcomes, showcasing the tangible value of applied AI.

Beyond my current role, I bring experience from DRDO, where I worked on applied ML for defense research, and Wipro, where I delivered enterprise-grade software solutions. These experiences gave me a strong foundation in problem-solving, innovation, and building robust systems.

I am passionate about creating AI products that deliver measurable results, with particular interest in conversational AI, speech AI, and applied research in LLMs.

Role
Machine Learning & ELM Engineer
Years of Experience
3 years

Skillsets

Natural Language Processing
Azure
AWS
MongoDB
Apache Kafka
Postgres
SQL
Scikit-learn
PyTorch
pandas
Python - 5 Years
MLFlow
LangGraph
LangChain
HuggingFace
Git
FastAPI
DVC
Docker
crewAI

Professional Summary

3Years

Jun, 2024 - Sep, 20251 yr 3 months
Machine Learning Engineer
Salesken.ai
Jan, 2024 - May, 2024 4 months
Research Intern
DRDO
Aug, 2021 - Jul, 2022 11 months
Project Engineer
Wipro

Applications & Tools Known

Microsoft Azure
MLflow
DVC
Git
Docker

Work History

3Years

Machine Learning Engineer

Salesken.ai

Jun, 2024 - Sep, 20251 yr 3 months

As a core member of the product development team at Salesken, I specialize in building and productionizing AI-driven systems to enhance customer interactions. My work involves leveraging Langchain, Large Language Models (LLMs), and state-of-the-art deep learning models to create robust, scalable solutions. Key Responsibilities & Accomplishments: Architected and deployed production-grade, high-performance LLM inference systems using vLLM, successfully serving over 5,000 requests per day, ensuring high availability and low latency for critical customer interaction analysis. Optimized and customized LLM's through advanced prompt engineering and by fine-tuning foundation models from Hugging Face on bespoke datasets, significantly improving performance for customer interaction analysis. These optimizations led to a 15% increase in model F1-score and a 20% reduction in inference time for key use cases.

Research Intern

DRDO

Jan, 2024 - May, 2024 4 months

Explored ELM, RDF, and Gaussian Probability Distribution for model parameter optimization, analyzing critical parameters to improve predictive accuracy and stability. Provided foundational insights for parameter estimation research.

Project Engineer

Wipro

Aug, 2021 - Jul, 2022 11 months

Spearheaded the development and deployment of Python Robot Framework and SQL automation solutions for Royal Sun Alliance (RSA), which successfully boosted operational efficiency by 50% and resulted in 200+ monthly hours saved. For Trane's Wipro Holmes Compliance Suite Deployment, I led the team in maintaining 99% uptime and successfully addressed 95% of all unforeseen issues, ensuring robust system performance.

Achievements

Flag Bearer Award, Wipro Limited
Student Council Representative, Christ University

Major Projects

3Projects

LLM-Agent Service with Custom Tools and Multi-Agent Collaboration

Developed a Langchain/LangGraph multi-agent system with OpenAI models for automated, personalized customer query handling. Integrated RAG for enhanced data retrieval and critic agents for response validation, leveraging historical data to deliver context-aware and adaptive interactions.

Scalable Multi-Node vLLM Deployment for Custom Model Inferencing

Deployed a multinode vLLM inferencing system with tensor parallelism and model sharding for low-latency, high-throughput processing of custom LLMs. Implemented auto-scaling and integrated with FastAPI/Langchain, achieving a 30% reduction in response time compared to standard Hugging Face pipelines.

GDPR Compliance for Customer Data at Salesken

Developed an automated GDPR compliance solution using Microsoft Presidio for PII detection and Hugging Face Transformers for advanced entity recognition, ensuring compliant handling of customer data across all stages. Automated detection and anonymization during data ingestion, storage, and sharing.

Education

Master of Science (Data Science)
Christ University (2024)
Bachelor's of Technology, Computer Science and Engineering
Jawaharlal Nehru Technological University (2021)

Certifications

Programming for everybody (getting started with python), university of michigan offered by coursera
Statistical inference using python, infosys springboard