Senior Machine Learning Engineer at Ctruh

Bengaluru, Karnataka, India

Bengaluru, Karnataka, India

Ahmedabad, Gujarat, India

Gurugram, Haryana, India







BSc, Mathematics
Sabarmati University
MicroMasters, Statistics and Data Science
MITx
Join Yuvraj and 10k+ other professionals on
Tal is the app where bosses meet engineers & builders directly, without middlemen
Download Tal BossSenior Machine Learning Engineer at Ctruh
I am a Machine Learning Engineer with around 5 years of experience architecting scalable, end-to-end AI products from concept to deployment. Driven by a builder's mindset, I have engineered four distinct production systems alongside my full-time work. My core expertise spans computer vision pipelines, multi-agent orchestration using LLMs and RAG, and comprehensive MLOps. Most recently, I led a three-person AI/ML R&D team at Styldod, where I developed complex 2D-to-3D scene reconstruction pipelines and launched an autonomous multi-agent Image AI system utilizing MCP architecture and LangChain. I am deeply passionate about optimization and infrastructure; I have successfully cut operational costs by 70% through deploying self-hosted LLMs and reduced inference latencies by over 85% to maximize API throughput and enable efficient GPU auto-scaling. Beyond enterprise roles, my independent engineering projects reflect a commitment to solving difficult AI challenges. I engineered ScaleWaveAI, a serverless AI inference platform with auto-scaling GPU clusters, and built KyoudaiClub, a full-stack real-time LLM character platform engineered for low-latency streaming. Equipped with a MicroMasters in Statistics and Data Science from MITx, I thrive at the intersection of deep research and production engineering, specializing in turning complex ML concepts into high-performing, user-centric applications.
Bengaluru
cursor
hermes agent
hugging face
langchain
langgraph
mcp
Senior Machine Learning Engineer at Ctruh, Jul 2026 - Sep 2026 · 3 mos, Bengaluru, Karnataka, India
Machine Learning Engineer at Styldod, Inc, Oct 2021 - Apr 2026 · 4 yrs 7 mos, Bengaluru, Karnataka, India, • Lead a 3 person AI/ML R&D team building computer vision, 3D, and generative AI solutions for visual automation. • Architected a 2D to 3D scene reconstruction pipeline using an ensemble of ML models and linear algebra for spatially accurate 3D furniture placement from single images. • Launched Reimagine Home's first Image AI Agent, orchestrated a multi-agent system featuring inter-agent communication. This collaborative architecture automated complex tool calling across image generation APIs, database fetching, live Amazon API integrations, and content moderation. • Engineered an automated content pipeline targeting AI search engines, driving a 17% increase in new user acquisition. Architected agentic tool calling to handle autonomous internet research, cross linking, and citation generation. • Cut operational costs by 70% by architecting a production MLOps system that deploys self-hosted LLMs and image generation models, serving millions of monthly requests. • Slashed VLM and image generation cold start latencies by over 85% (down to 50s and 37s) by optimizing inference pipelines, maximizing API throughput and enabling efficient GPU auto-scaling. • Trained Diffusion models across multi GPU clusters on millions of images. Built a custom Human-in-the-Loop evaluation pipeline to bypass unreliable loss graphs and ensure high fidelity production deployments.
Research Engineer at RAx Labs, Jul 2021 - Oct 2021 · 4 mos, Ahmedabad, Gujarat, India, • Developed an NLP pipeline and Twitter bot for real-time research paper summarization, enhancing algorithms for sentence ranking and information extraction. • Deployed scalable backend architectures via Docker on AWS ECS with a full observability stack (Prometheus, Grafana, DataDog) for production monitoring. • Optimized heavy Elasticsearch queries indexing thousands of complex PDFs, minimizing retrieval latency for academic research data.
AI/ML Intern at Zimyo, Mar 2021 - Jun 2021 · 4 mos, Gurugram, Haryana, India, • Created a conversational Chat Bot it detect the sentiment and take actions accordingly. It’s also capable of correcting user errors out of the box. • Tensorflow, NLTK is used to train the model. • NGINX is used to handle to requests. • AWS ec2 used for deployment.
BSc, Mathematics, Sabarmati University
MicroMasters, Statistics and Data Science, MITx
yuvraj97.com: https://www.yuvraj97.com/