How LLM Inference Got 10x Cheaper
How the field slashed LLM inference costs through attention redesigns, IO-aware computation, KV cache paging, and speculative decoding, and what comes next.
TransformersLLM Inference
Engineering Manager • Frontend • AI/ML
Siddharth Parmar (Sid)
Building intelligent user experiences at the intersection of frontend engineering and AI. Specializing in React, TypeScript, and LLM-powered applications.
About
I'm an Engineering Manager who lives at the intersection of frontend craft and AI. I build fast, polished React and TypeScript applications — and then make them intelligent with LLMs, agentic systems, and production AI tooling.
At Quo (formerly OpenPhone), I lead frontend and AI engineering across the product — shipping production UI alongside agentic features like multi-agent orchestration, MCP tooling, and end‑to‑end intelligent workflows. I'm most energized when a project demands both pixel‑level attention and deep technical problem‑solving.
Outside of shipping code, I mentor developers, write about what I'm learning on my blog, and dig into whatever emerging technology catches my curiosity. I believe the best engineering teams are built on honest code reviews, shared context, and a genuine interest in each other's growth.
If you're building something interesting — whether it's a frontend‑heavy product, an AI integration, or something in between — I'd love to hear about it at siddharthparmar7@gmail.com.
Latest Writing
AI Agents, MCP Tools, and Skills: Toward Smarter Workflows
Why naive MCP usage breaks at scale, how code execution fixes context bloat, and how skills turn agent workflows into reusable building blocks.
AI AgentsMCP tools
Experience
Quo (formerly OpenPhone)
Dec 2023 — PresentEngineering Manager
Jan 2026 — Present
Senior Software Engineer — Frontend • AI/ML
Dec 2023 — Dec 2025
- Lead frontend development and AI engineering across the product — from React/TypeScript UI to LLM-powered features
- Design and build agentic systems: multi-agent orchestration, MCP tooling, A2A protocol, evaluation loops, and reusable agent skills
- Own AI tooling and infrastructure — context engineering, model optimization (distillation, quantization), and production inference pipelines
- Ship end-to-end intelligent workflows that connect frontend surfaces to agentic backends
Calabrio
Dec 2021 — Dec 2023Senior Software Engineer
- Spearheaded integration of micro front-end architecture with React tech stack, boosting scalability and productivity by 50%+
- Developed UI component library with React, TypeScript, Storybook, Material-UI, meeting a11y and i18n standards (60% productivity increase)
- Led architectural decisions for full-stack technologies, enhancing software development experience
- Implemented automated tasks using Jenkins, AWS, Docker, and Azure, reducing single point failures by 70%+
- Mentored interns in growth-oriented environment, resulting in 20% team productivity increase
CareCru
Apr 2020 — Dec 2021Software Engineer
- Designed and deployed Micro Front End services using React, Redux, TypeScript, GraphQL, Apollo (20% user satisfaction increase)
- Spearheaded migration from legacy codebase to microservice architecture, enhancing scalability and testability by 40%
- Contributed to React-based UI component library, influencing key design decisions (30% code reusability, 50% productivity boost)
RED Academy
Jan 2018 — Apr 2020Lead App Developer
- Developed and trained teams in Full-Stack development using JavaScript, React, React Native, GraphQL, Redux, Jest
- Led technical stack including NodeJS, ExpressJS, Apollo, REST APIs, Postgres, MongoDB
- Mentored teams and guided them to achieve success in full-stack development
North Island College
Jul 2017 — Dec 2017Instructor
- Taught PHP and web development courses
- Set up secure AWS server for hosting
- Mentored students in web development fundamentals
Skills
Software Development
JavaScriptTypeScriptPythonReactReact HooksReduxApolloNodeJSGraphQLREST APIsHTML/CSSCSS-in-JSSASSGridFlexboxReact Testing LibraryJestStorybookWebpackVercelGitPostgresMongoDBAWSCI/CDJenkinsMicro Frontends
AI/ML Engineering
LangChainHugging FacePyTorchTensorFlowRAGAgentsFAISSFine-tuningPrompt EngineeringLLMsTransformersSupervised LearningUnsupervised Learning
Certifications
Unsupervised Machine Learning
University of Waterloo
Issued Feb 2025
Machine LearningClusteringDimensionality ReductionPattern RecognitionFeature EngineeringData Mining
Foundations of Large Language Models: Tools, Techniques, and Applications
University of Waterloo
Issued Nov 2024
Large Language ModelsNatural Language ProcessingDeep LearningAI ApplicationsModel TrainingFine-tuningText Generation
AlgoExpert Certificate
AlgoExpert
Issued Oct 2022
Credential ID: AE-5c2bad4525
AlgorithmsData StructuresCoding InterviewsTime ComplexitySpace ComplexityProblem Solving
The Data Science Course 2022: Complete Data Science Bootcamp
Udemy
Issued Jun 2022
Credential ID: UC-1cd7c4b9-3086-431f-a6db-115dcfe3cb08
PythonData AnalysisMachine LearningData VisualizationStatistical ModelingPandasNumPy
Statistics for Data Science and Business Analysis
Udemy
Issued Jul 2021
Credential ID: UC-874b7bb7-b1f4-4fb8-a5c1-a9b899b284ad
StatisticsProbabilityHypothesis TestingRegression AnalysisData AnalysisBusiness Analytics
Hootsuite Certified Professional
Hootsuite
Issued Mar 2016
Social Media MarketingContent StrategyDigital Marketing
Professional Contact / Resume
Business Inquiries (Sid/Siddharth): siddharthparmar7@gmail.com
For professional opportunities, recruitment, technical collaboration, and business partnerships only.
GitHub (opens in new tab) · LinkedIn (opens in new tab) · Twitter/X (opens in new tab) · Medium (opens in new tab)