Headline

AI/ML Engineer | LLM Pipelines · Production ML Systems · Feature Engineering | Open to ML Engineering & AI Roles

About

I build AI/ML systems that ship to production — from LLM pipelines to feature-engineered ML models serving real workloads. At Netlink Software Group, I improved ML model accuracy by 12% on 300K+ records, cut API latency by 30% on a high-throughput inference pipeline, built an LLM system with GPT-4 and LangChain that reduced analyst effort by 40% per cycle, and engineered an AI-powered calling system handling 4,000–5,000 calls/day — orchestrated via n8n, integrated with HubSpot CRM, and powered by OpenMic for real-time voice AI. My stack: PyTorch, scikit-learn, Hugging Face Transformers, LangChain, OpenAI/Claude APIs, FastAPI, AWS (EC2, S3, CloudFormation), MongoDB, Docker, n8n, HubSpot, OpenMic.

Location

Bhopal

AI stack

anthropic apis

claude code

claude

codex

cursor

hugging face

Work experience

Software Engineer Intern at Netlink Software Group America Inc, Nov 2025 - Present · 11 mos, Bhopal, Madhya Pradesh, India, Working on production ML and backend systems for a US-based B2B technology services firm serving enterprise healthcare and supply chain clients. - Reduced API response latency by 30% by refactoring database query patterns and restructuring backend architecture across a high-throughput ML serving pipeline - directly improving SLA compliance for enterprise clients. - Improved ML model accuracy by 12 percentage points through advanced feature engineering and preprocessing on 300,000+ structured project records; stabilized production performance with 5-fold cross-validation and Bayesian hyperparameter tuning. - Architected an LLM-powered tender analysis system using GPT-4 and LangChain that automated requirement extraction and proposal generation, eliminating 40% of analyst manual effort - saving~12 analyst-hours per procurement cycle. - Engineered a real-time vehicle density estimation module using deep learning computer vision for adaptive signal timing; integrated via REST API into a cloud-deployed traffic optimization system.

AWS Solutions Architect at ETHNUS, Jan 2025 - Apr 2025 · 4 mos, Designed and deployed cloud infrastructure for enterprise clients as part of Ethnus's AWS training and consulting program. - Architected and deployed AWS infrastructure (EC2, S3, IAM) using CloudFormation infrastructure-as-code templates across dev/staging/prod environments, reducing provisioning time by ~60% and eliminating configuration drift. - Implemented Auto Scaling groups and Elastic Load Balancing to deliver multi-AZ fault tolerance and reduce monthly AWS EC2 spend through demand-based right-sizing.

Data Specialist at IIM Mumbai, Jan 2025 - Mar 2025 · 3 mos, • Built NLP pipelines using Python and Hugging Face Transformers (BERT-based classification) to automate business intelligence report generation from unstructured data. • Developed LangChain-based prototypes integrating OpenAI APIs to produce structured insights from raw business datasets, reducing analyst reporting time. • Designed AI-assisted recommendation frameworks using Pandas and AWS S3, delivering personalized decision support outputs.

Science Specialist at India Space Lab ( भारत अंतरिक्ष प्रयोगशाला ), Dec 2024 - Jan 2025 · 2 mos, Delhi, India, Developed and implemented Python and Django modules for drone and satellite programs, improving data management and enabling successful prototype testing that advanced the team’s research and space entrepreneurship projects. And made specialized modules in Advanced Drone Technology, CanSat and BLS CubeSat Satellite Programs, Astronomy, and Space Entrepreneurship.

Python Developer at YBI Foundation, Sep 2024 - Oct 2024 · 2 mos, Chennai, Tamil Nadu, India

Education

Bachelor of Technology, Computer Science, VIT Bhopal University

sagar public school

Public links

github: https://github.com/manavp7

manav-portfolio-cyan.vercel.app: https://manav-portfolio-cyan.vercel.app/