Open to remote US roles

I build usefulAI systems thatactually ship.

Satvik Mishra is a Full Stack & Applied AI Engineer shipping production-ready LLM systems, RAG pipelines, and scalable backend architecture for real teams and real products.

Start a conversation →
Scroll for the proof

Engineering with
measurable impact.

Selected outcomes / 2025–26
0%

Cut in LLM pipeline operating costs via caching, batching & model routing

0+

Production systems shipped across 8 domains — agents, RAG, SaaS & NLP tooling

0+

Concurrent users handled by backend design optimizations with zero-downtime feeds

0%

Contextual retrieval accuracy across production RAG systems

The working toolkit

Breadth, without
the buzzwords.

From data model to deployment, this is the stack used to design, build, evaluate, and ship intelligent products end to end.

01

Languages

PythonTypeScriptJavaScript (ES6+)JavaC++SQL
02

AI / ML Engineering

RAG Pipelines Vector Embeddings Agentic Workflows Autonomous Reasoning Agents Tool Calling & Structured Outputs ASR / Speech-to-Text TTS / Speech Synthesis Real-Time Voice Agents Conversation State & Guardrails Fine-Tuning Gemini / LLaMA / GPT
03

Backend & Frameworks

FastAPINode.jsExpress.jsReact.jsViteTailwind CSSREST APIs
04

Databases & Vector Stores

MongoDBPostgreSQLMySQLChromaDBQdrant
05

Infra & Tooling

AWS (EC2, S3)DockerGitHub ActionsCI/CDVercelRenderNginx
06

System Design & Collaboration

System DesignAsync CollaborationCode ReviewNotionJiraGit / GitHub
Career timeline

Built in the
real world.

Applied AI work, product ownership, live systems, distributed collaboration, and measurable outcomes.

AntarikshTV

Remote (India)

Jan 2026 — Present

AI Engineer

  • Architected a content intelligence system using LangChain, Gemini, and ChromaDB that analyzes 500+ historical videos and live news signals to surface high-potential topics, contributing to a 30–35% lift in average video engagement measured over 90+ days of production usage.
  • Built automated multi-step research agents for a 10-person writing team to retrieve insights from internal video archives, reducing manual research effort and improving script delivery throughput.
  • Designed and scaled FastAPI backend services and async data pipelines supporting 50+ concurrent internal users, maintaining real-time analytics availability and zero-downtime topic recommendation feeds.
  • Leading fine-tuning of a domain-specific scriptwriting LLM on 3+ years of proprietary scripts, targeting faster content generation and reduced production overhead.
  • Collaborate with a fully distributed team using GitHub, Notion, and Slack, maintaining documentation and code reviews for all shipped features in an async-first workflow.
AntarikshTV

Remote (India)

Nov 2025 — Dec 2025

AI Engineer Intern

  • Built an internal AI chatbot using LangChain and ChromaDB to query organizational knowledge bases and surface real-time project status and content pipeline insights for the production team.
  • Developed a centralized task-tracking platform for content moderation workflows, improving cross-team visibility and contributing to a ∼15% improvement in tracked team output over the 2-month engagement.
  • Shipped internal coordination tooling to streamline communication across a 50+ member distributed production team, establishing async engineering practices from day one.
RBLottery

Remote (India)

Mar 2025 — Sep 2025

Co-Founder & Founding Engineer

  • Co-founded and sole-engineered a full-stack digital lottery platform from zero using the MERN stack (MongoDB, Express, React, Node.js), JWT authentication, role-based access control, and Tailwind CSS.
  • Scaled the platform to 6,300+ registered users with a 78% repeat usage rate, reaching profitability within months as the only engineer on the team.
  • Architected REST APIs with Express.js and MongoDB, implemented admin/user/moderator access control, and deployed production infrastructure on Hostinger and Netlify with 99%+ uptime.
  • Shut down in September 2025 following the Government of India’s Online Gaming Regulation Bill (August 2025), which banned all forms of digital lottery nationwide — not due to product or business failure.
Independent

Remote (India)

Apr 2024 — Feb 2025

Contract Software Engineer

  • Designed, built, and delivered 23 production applications from ambiguous idea to live deployment for clients across 8 sectors, owning discovery, architecture, implementation, testing, infrastructure, and handover end to end.
  • Engineered customer-facing AI voice agents with real-time speech-to-text, text-to-speech, LLM orchestration, tool calling, conversation-state management, and fallback flows—turning natural-language calls into reliable operational workflows.
  • Built autonomous reasoning-agent systems for high-accuracy environments, combining multi-step planning, retrieval-augmented generation, structured outputs, verification loops, and human-in-the-loop escalation paths.
  • Developed scalable full-stack products and backend services using React, Node.js, FastAPI, Python, MongoDB, PostgreSQL, and REST APIs, translating client requirements into maintainable production systems.
  • Designed RAG pipelines with document ingestion, chunking, embedding retrieval, reranking, and source-grounded responses to improve answer relevance and reduce hallucinations in domain-specific knowledge workflows.
  • Implemented secure deployments, authentication, role-based access control, monitoring, and CI/CD workflows across cloud-hosted applications, enabling dependable releases without sacrificing delivery speed.
Selected systems

Things I’ve
brought to life.

Five featured systems across multi-agent AI, semantic search, NLP infrastructure, workflow design, and content intelligence.

CapTon AI application screenshot
01 / Multi-Agent Content Caption System

CapTon AI

Production-ready multi-agent platform using CrewAI that converts long-form video scripts into SEO-optimized assets via coordinated analyzer and writer agents.

CrewAILangChainPython
Open live project ↗
M.A.D.H.A.V. application screenshot
02 / Vectorized Semantic Search

M.A.D.H.A.V.

Retrieval-augmented AI assistant that vectorizes 700+ Bhagavad Gita verses into ChromaDB, achieving 95% contextual retrieval accuracy against a benchmark set.

FastAPIChromaDBRAG
Open live project ↗
03 / Legacy Hindi Encoding → Unicode

HinLegacy

Python package for detecting and converting legacy Hindi font encodings such as Kruti Dev, DevLys, and Chanakya into Unicode-ready text for modern NLP, RAG, and AI pipelines.

PythonNLPPyPI
View on PyPI ↗
VectorFlow UI screenshot
04 / Flowchart Agent UI Recreation

VectorFlow

A stylized UI frontend-only recreation of VectorShift and n8n with a DAG detection protocol inside of it.

ReactFlowLucideReact
Open live project ↗
ContelliSense application screenshot
05 / AI-Powered YouTube Content Pipeline

ContelliSense

Content Intelligence System that automates the entire content creation pipeline for YouTube, from research to data trend analysis to pipeline streamlining, using a multi-agent architecture.

Next.jsLangGraphPython
Launching soon
The next chapter

Have a hard problem?
Let’s ship it.

codewithsatvik@gmail.com →