Shubham Agrawal — Product Manager

📌 "Data is the new oil, and I'm here to drill it." 👋 Hi, I am a Full-Stack ML Engineer with extensive experience in machine learning and data engineering. I specialize in architecting, developing, and deploying scalable AI solutions. My experience spans across end-to-end analytics projects, encompassing Data Analysis, Data Engineering, Data Visualization, Machine Learning, Model Deployment, and developing comprehensive analytics frameworks to solve complex business challenges. Currently, I built a recommendation service for a short video platform. I also developed a sales chatbot, a sales call summarizer and a telecaller assist bot integrated with partner CRM. Previously at Aspire FinTech, I reduced credit risk with defaulter detection algorithms and engineered a decision/rules engine. I also created LLM-based solutions for customer interaction. Previously at Siksha Sopan, I developed the backend infrastructure of the EdTech platform. ✅ Programming: Python (Numpy, Pandas, NLTK, Scikit-Learn, Matplotlib), PyTorch, HuggingFace, Keras, SQL, C. ✅ Machine Learning: Logistic Regression, Random Forest, XGBoost, kNN Classifier, K-means Clustering, Feature Extraction (PCA, Factor Analysis), Neural Networks, Natural Language Processing (Text Analytics – PII Extraction), Large Language Model(LLM}, Convolutional Neural Network. ✅ Maths: Experimental Design, Hypothesis Testing (A/B Testing), Linear Algebra, Statistics, Probability ✅ Deployment & DevOps: AWS, PyCharm, Git, Jira, Docker, Linux ✅ Databases: PostgreSQL, MySQL, Redis, DynamoDB, MongoDB, Clickhouse ✅ Backend Frameworks: FastAPI, Django ✅ Business Domain Expertise: Credit-risk Analysis, Content Recommendation, AI sales, EdTech 🤝 Let's connect to discuss innovative AI solutions!

Stackforce AI infers this person is a Full-Stack ML Engineer specializing in Fintech and EdTech industries.

Location: Gurugram, Haryana, India

Experience: 5 yrs 9 mos

Skills

Conversational Ai
Data Engineering
Machine Learning
Backend Development
Data Science
Cloud Services

Career Highlights

Expert in architecting scalable AI solutions.
Led teams to deliver production-grade AI features.
Significant contributions to credit risk reduction.

Work Experience

VerbaFlo.AI

Senior AI Engineer (1 yr 5 mos)

Stealth Startup

Artificial Intelligence(AI) Scientist (1 yr 1 mo)

Shiksha Sopan - IIT Kanpur

Developer (11 mos)

Aspire UPI Pay Later

Data Scientist (2 yrs 4 mos)

Alumnus Software Limited

Data Scientist (11 mos)

Education

Entrepreneurship at Y Combinator

Bachelor of Technology - BTech at Heritage Institute of Technology

6-10 Std. at Gita Niketan Awasiya Vidyalaya, Kurukshetra

Shubham Agrawal

Product Manager

Gurugram, Haryana, India5 yrs 9 mos experience

AI EnabledAI ML Practitioner

Key Highlights

Expert in architecting scalable AI solutions.
Led teams to deliver production-grade AI features.
Significant contributions to credit risk reduction.

Stackforce AI infers this person is a Full-Stack ML Engineer specializing in Fintech and EdTech industries.

Contact

Skills

Core Skills

Conversational AiData EngineeringMachine LearningBackend DevelopmentData ScienceCloud Services

Other Skills

AI AgentsAI PromptingAWS DeploymentAgentic AIAlgorithmsAlgorithms & Data StructuresAmazon Web Services (AWS)Apache SparkApache Spark MLAsking the Right QuestionsAsynchronous ProgrammingBig DataBusiness AnalysisC++Cascading Style Sheets (CSS)

About

Experience

5 yrs 9 mos

Total Experience

1 yr 5 mos

Average Tenure

1 yr 5 mos

Current Experience

Verbaflo.ai

Senior AI Engineer

Jan 2025 – Present · 1 yr 5 mos · United Kingdom · On-site

Architected a multi-modal Conversational AI Agentic system for Calls, Chat Widget, WhatsApp, and Email, serving diverse B2B United Kingdom based accommodation providers.
Led a 10-member AI engineering team to deliver production-grade features including lead acquisition, customer support, campaigns, data extraction, conversation summarization, FAQ answering via RAG, and dynamic property recommendations.
Applied advanced prompting techniques (CoT, one-shot, few-shot) and designed Mermaid-based conversational flows for structured interactions.
Optimized RAG pipeline accuracy and cut response latency to 1 second, improving user experience.
Implemented jinja based Prompt pruning, Conversational Context Management & Engineering to reduce tokens, costs, latency, and prevent length-finish errors.
Built a Prompt Management System with versioning, DB updates, and cross-modality compatibility, integrated with a Property Management System for real-time room recommendations.

Conversational AIData ExtractionConversation SummarizationDynamic Property RecommendationsPrompt EngineeringData Engineering

Stealth startup

Artificial Intelligence(AI) Scientist

Dec 2023 – Jan 2025 · 1 yr 1 mo · United States

Architected Recommendation Service from scratch for our short video platform & sales-assist Startup.
Achieved ultra-low latency by leveraging asynchronous programming and FastAPI, resulting in a 2x reduction in latency.
Engineered sales assist bot, integrated with partner CRM to assist telecallers during live calls, enhancing customer engagement by 10% and sales by 5%.
Created sales chatbot with features like token streaming, next question suggestions and video recommendations, within a week.
Launched sales call summarizer, cutting call review time by half.
Created dashboards using SQL in Metabase & Redash to track product metrics, content performance etc. which led to faster product iteration and decision-making.
Tech stack: FastAPI, Clickhouse, Postgres, DynamoDB, Grafana, Kafka, Docker, Deepgram, OpenAI, MongoDB, VectorDB.

Recommendation ServiceAsynchronous ProgrammingSales ChatbotSQL DashboardsFastAPIMachine Learning+1

Shiksha sopan - iit kanpur

Developer

Apr 2022 – Mar 2023 · 11 mos · Remote

EdTech Non-Profit Organization led by Prof H.C. Verma (IIT Kanpur)
Developed the backend infrastructure for the registration and examination portal using Django.
Designed and implemented RESTful APIs and database models in Django to manage user registration and exam data.
Utilized Python, Django, and SQL for backend development, ensuring seamless data management and retrieval.
Leveraged AWS for deployment and cloud services, ensuring high availability and scalability.
Employed version control with Git to manage and collaborate on the codebase.

Backend DevelopmentRESTful APIsDatabase ManagementAWS DeploymentCloud Services

Aspire upi pay later

Data Scientist

Aug 2021 – Dec 2023 · 2 yrs 4 mos · Bengaluru, Karnataka, India

Credit Risk | Lending | UPI Pay Later | FinTech Startup
Designed algorithm to identify defaulter SMS, leading to a significant reduction in credit-risk.
Trained ML-models for defaulter identification & developed a fuzzy name-matching algorithm.
Engineered Decision/Rules Engine from scratch using Django, enabling the deployment, simulation & A/B testing of data pipelines, model pipelines and risk policies. This service personalized product assignments, leading to lower risk, reduced onboarding costs, and higher approval rates.
Developed LLM sales chatbot to interact with customers on WhatsApp.
Implemented Information-extraction from SMS using LLM.
Mentored the DS team in integrating data and model pipelines into the production codebase, enhancing productivity.
Data: SMS(text), Bureau(JSON), KYC(text, image, video).
Models: Logistic Regression, Decision Trees, Random Forest, Gradient Boosting.
Tech Stack: Python, SQL, Django, AWS, Pandas, Numpy, scikit-learn, OpenAI, Sagemaker.