Sangam Man Buddhacharya
ML Engineer @ Fetch | AI Agents to Simplify and Automate Tasks | ML Engineer & Data Scientist | LangGraph LangChain, Tensorflow, PyTorch | Helped 8+ companies solve their business problem using data
- Role
- Machine Learning Engineer at Fetch
- Location
- Portland, OR, US
- LinkedIn followers
- 500 followers
About Sangam Man Buddhacharya
I am a Machine Learning Engineer with over 4 years of hands-on experience in building…
Experience
Machine Learning Engineer
Jun 2025 — Present · Chicago, IL, US
At Fetch, I drive the design and scaling of machine learning systems that power personalization, relevance, and ranking across our consumer platform. My work focuses on building robust ML infrastructure, leading architecture and technical direction. Develop zero-to-one real-time Machine Learning systems and data pipelines.Key Achievements:• Designed and implemented a multi-agent LLM system in LangGraph to generate high-fidelity product catalog descriptions from UPC barcode details, reducing duplicate descriptions from 3M to 1M and improving catalog fidelity scores by 15%.• Built an agentic AI to automatically extract key product attributes (e.g, age group: infant, toddler, adult, senior; form: liquid, gel, cream, spray, powder) from catalog product description, cutting human review time by 70% and boosting B2B recommendations by 20%.• Developed a Scraper AI agent using LangChain and GPT-4.1 to extract product metadata (description, brand, manufacturer, category) from barcode lookup websites and google search engine, reducing third-party API costs from $10K/month to $1.5K.• Deployed a scalable LLM-powered description generation service on Amazon EKS, capable of handling 100+ RPS, and successfully generated high-quality descriptions in under 30 minutes.• Automated the product description generation pipeline for millions of low-fidelity catalog items by integrating AWS Lambda (with S3 triggers), webhooks, and n8n workflows — reducing manual effort by 40% and significantly improving scalability.• Built and implemented a CI/CD pipeline in Bitbucket to automate code linting, error detection, and bug prevention; streamlined Docker image creation and deployment to AWS ECR and ECS, improving code testing efficiency by 20%.
Education
Oregon State University
Dual Master's degree
2023
Tribhuvan University, IOE, Pulchowk Campus
Bachelor's degree
2016 — 2021
Find verified contacts for anyone on LinkedIn
Unifers gives sales teams verified emails and direct dials, enriched profiles, and outreach that lands in the inbox.
Free plan included · No credit card required
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.