Saahil Basak

Saahil Basak

Software Engineer 1

Delhi NCRSoftware Engineering
View LinkedIn View resume
5roles
100skills
2education
3credentials

About

Engineer who repeatedly takes ambiguous problems, chooses an architecture, ships it, and owns the result.

Experience

Software Engineer 1 — E-commerce Startup

ECOWELL Health & Beauty Pvt Ltd

Dec 2025 — July 2026 · New Delhi, India

Worked with marketing and product teams to turn business goals into production-ready full-stack and agentic systems, with ongoing site support across six properties. Delivered full-stack and agentic AI solutions for the marketing team using async crawling, parallel specialist agents, and structured JSON contracts, reducing recurring workflows from weeks to hours. Maintained and optimized six Shopify and React websites, improving LCP, asset optimization, and atomic component systems; achieved 1.5-second LCP on a media-heavy site and reduced Shopify load time by 86%.

Software Engineer 1 — Founding Team — Edtech Startup

MEDJEEX Edutech Pvt Ltd

Sep 2024 — Nov 2025 · Gurugram, India

One of two founding full-stack engineers on a five-person team; evolved the platform from a single-service MVP into a 10-service, AI-augmented LMS serving 2,000+ students. Migrated a monolith into 10 domain-oriented microservices across four independently deployable repositories, with Controller–Service–Repository layering and Redis caching. Built a WebRTC classroom on a MediaSoup SFU with role-based media routing, bandwidth-aware stream forwarding, moderator controls, and server-side recording. Orchestrated asynchronous LangChain/LangGraph workflows through BullMQ to process 2,000+ lecture transcripts into notes, flashcards, MCQs, and podcasts using the Gemini API, plus a lecture-timestamp-grounded AI tutor. Designed trust-tiered assessment execution models scaling to 2.5k+ concurrent test-takers. Built offline-first analytics and gamification systems, a knowledge forum with hybrid feed ranking and keyword/vector search, a Python/OpenCV OMR engine with 99.8% accuracy and 80% faster grading, a FastAPI verification dashboard, a Word-to-test ETL pipeline, and centralized field-level RBAC for internal operations tooling.

Software Engineer

ECOWELL HEALTH & BEAUTY PVT.LTD

9 mos · Delhi, India

Led all technical systems across a multi-brand group — web, commerce, analytics, and growth — as the sole engineer. • Engineered a 14-phase automated AI SEO pipeline transforming raw website crawls into prioritized, evidence-backed strategy playbooks — leveraging Crawl4AI, LiteLLM, and a multi-agent orchestration pattern (E-E-A-T, SERP, Entity, Psychology) to reduce manual audit time from weeks to hours. • Built a suite of internal tools and automations to handle daily business operations, including custom pixel tracking, ad script automation (Meta, Amazon), and crawlability fixes — replacing high-cost 3rd-party Shopify apps with in-house technical components to eliminate recurring overhead. • Built a React enterprise website (Jayshree International) end-to-end using an atomic design system with centralized theming and data config — achieved 0.4s LCP on an image-and-video-heavy page. • Reduced LCP from 12s to 1.7s on media-heavy Shopify storefronts through lazy loading, asset optimization, and delivery strategy. • Implemented multi-platform analytics and ad-tracking (Google Analytics, Microsoft Clarity, Meta Pixel) across brands, enabling data-driven growth decisions. Stack: Node.js, React, Agentic AI, Shopify (Liquid), Crawl4AI, LiteLLM, Google Analytics, Atomic Design, Core Web Vitals

Software Engineer

MEDJEEX

1 yr 1 mo · New Delhi

Built an AI-powered edtech platform entirely from scratch — full product, full stack, independently owned across backend and frontend. • Built a complete edtech platform — 30+ screens in React/Redux, Node.js/MongoDB backend — covering authentication, test engine, course delivery, results analytics, and AI-assisted learning. Load-tested to 1,000 concurrent users with API responses consistently under 180ms. • Architected a service-separated backend deploying independent services (auth, tests, courses, AI, analytics) within AWS free-tier constraints. Built a live class system replacing paid Zoom integration — instructor video prioritization, teacher controls, live comments, and in-class polls. • Built a LangChain + Google AI pipeline for syllabus-scoped Q&A, answer explanations, AI-generated notes, flashcards, and interactive voice Q&A. Validated outputs against a structured test suite covering factual accuracy and content safety. • Built a custom behavioral analytics pipeline (replacing Google Analytics) modeling engagement metrics — streaks, watch time, playback speed, quiz performance, AI usage — with client-side buffering and batched sync. • Built a community discussion platform with syllabus-aware auto-tagging, AI-generated first responses, threaded discussions, and a custom ranking algorithm. Stack: Node.js, React, Redux, MongoDB, LangChain, Google AI, JWT, Razorpay, AWS

Machine Learning Intern

CodSoft

4 mos · India

Worked on end‑to‑end ML pipelines over real‑world data (movie metadata, SMS logs, card transactions) for text classification and fraud/anomaly detection. • Built NLP‑based classifiers for movie genres and spam SMS. • Developed fraud detection models on transaction data tuned for high recall. Stack: Python, Machine Learning, Deep Learning, NLP, scikit‑learn, pandas, NumPy, Jupyter, Classification, Anomaly Detection, Fraud Detection, Spam Detection, Logistic Regression, SVM, Random Forest, Isolation Forest, One‑Class SVM, LSTM, CNN, Git

Education

Cooch Behar Government Engineering College

Bachelor of Technology, Computer Science & Engineering

Jul 2020 — Jul 2024

CGPA: 8.75

MAKAUT (WB)

Bachelor of Technology, Computer Engineering

CGPA 8.75

Skills

Node.jsTypeScriptJavaScriptPythonReal-Time SystemsAI/Agentic SystemsMicroservicesEvent-Driven ArchitectureExpressNestJSFastAPIREST APIsGraphQLSocket.IOJWTOAuthAI AgentsLangChainRAGVector DatabasesWebRTCMongoDBPostgreSQLRedisMySQLGitHubAWSAmazon EC2Amazon S3AWS LambdaGCPDockerCI/CDReactReduxTailwind CSSLLM IntegrationPerformance OptimizationCore Web VitalsShopifyAtomic DesignMediaSoupLangChain/LangGraphGemini APIBullMQOpenCVOMRIndexedDBMongoDB Vector SearchFull-Stack DevelopmentMarket ResearchUser Experience (UX)Marketing StrategyAttention to DetailResponsible AIAgentic AI DevelopmentReal-time WebReal-time SimulationServer AdministrationReal-time Control SystemsServiceStackWorkflow EnginesMapping SoftwareReal-time RenderingContent FilteringCode RefactoringApp Store OptimizationWeb LanguagesConversational AIBrowserStackStabilizationSite LayoutMonetizationDatabase QueriesProblem SolvingCommunicationLeadershipComputer ScienceMarkup LanguagesWAMPDocker ProductsCore MLPython (Programming Language)React.jsAmazon Web Services (AWS)Google AnalyticsMATLABPyTorchTensorFlowKaggleMicrosoft ExcelC (Programming Language)C++Web PerformanceSearch Engine Optimization (SEO)Product DevelopmentSoftware ArchitectureAI/MLProcess Flow DocumentationSupport Vector Machine (SVM)

Projects

High-Speed Market Data Ingestion Pipeline for Matlab ( FREELANCE CLIENT )

Engineered a live market data pipeline for a hedge fund's NSE futures trading system. • Designed and deployed a decoupled ingestion pipeline bridging live NSE futures market feeds into a MATLAB execution environment via a Python WebSocket-to-TCP bridge — keeping the trading system's interface unchanged. • Built binary serialization (MessagePack) and auto-reconnection for data integrity and feed resilience under sustained throughput across hundreds of symbols. • Implemented in-memory O(1) snapshot access for sub-millisecond reads by the trading system, integrated with MATLAB's parallel processing for concurrent symbol handling. • Validated across millions of market ticks in live trading conditions with zero data loss under sustained load. Stack: Python, MATLAB, WebSockets, TCP/IP, MessagePack, Parallel Computing