Skip to content
vikash.kumar

About

I build AI systems that survive contact with production — not demos.

Software engineer with 5+ years across the full stack, now focused on generative AI and agentic systems. At Armor I work on a multi-agent AI platform for security operations: LLM orchestration, agentic RAG over security telemetry, tool and MCP integrations, and AI-generated attack-story timelines that turn a pile of raw detections into a narrative an analyst can actually read.

Most of my time goes to the parts that decide whether AI actually ships: retrieval quality, evaluation and observability, guardrails, failure recovery, and cost. The model is rarely the hard part — the distributed system around it is.

Before that I shipped full-stack products at Volkswagen Group Technology Solutions, Adeption, and BigBinary — React, Node and TypeScript front to back, owning architecture, testing, and delivery.

I also ship outside work: this site, a book-sharing platform, and a shelf of published Android games — all live, all built end to end. Shipping small things alone is the fastest way I know to stay honest about what actually works.

LinkedIn →GitHub →

Experience

  1. Software Developer II · Armor Defense

    Mar 2025 — present

    Pune, India

    Building an AI/agentic platform for security operations, plus the full-stack product surface around it.

    • Design and ship multi-agent workflows (planner / worker / reviewer roles) that run long-horizon security tasks end to end, with failure recovery, requeue and timeout handling, and heartbeat-based liveness so runs don't silently stall.
    • Built agentic RAG over security telemetry and product documentation: connector ingestion, chunking, embeddings and vector retrieval, wired into both chat and agent tool calls.
    • AI-generated attack stories: a pipeline that correlates detections and alerts into an analyst-readable incident timeline, cutting triage reading time for the SOC.
    • Expose internal systems to agents as typed tools over MCP, and build the chat, build and workflow surfaces engineers and analysts use day to day.
    • Own the reliability side: prompt and role design, regression gates on agent behaviour, observability and cost control.
    • Work in a security-first, data-sovereign environment — isolated execution, least-privilege access, and no customer data leaving its boundary.

  2. Software Engineer · Volkswagen Group Technology Solutions India

    Oct 2023 — Mar 2025

    Pune, India

    Full-stack engineering on Volkswagen Group's internal engineering and delivery platforms.

    • Innovation Award (Sep 2024) for a Java migration tool that automated moving legacy applications onto modern frameworks, removing large amounts of manual rework and migration error.
    • Debutant Award (Mar 2024) for early impact in the team.
    • Grew from full-stack developer into dev-lead responsibilities: technical direction, architecture decisions, code review and delivery ownership.
    • Built and maintained React front ends and Node/Java services with GraphQL and REST APIs, test-driven throughout.

  3. UI Developer · Adeption

    Jul 2022 — Sep 2023

    Pune, India

    Front-end and full-stack development on a SaaS leadership-development platform.

    • Built the product UI in React and TypeScript, including a reusable SCSS design-system layer used across the app.
    • Implemented real-time collaboration features over WebSockets.
    • Integrated authentication and SSO with OpenID Connect (OIDC).

  4. Software Engineer · BigBinary

    Oct 2020 — Jun 2022

    Remote

    Full-stack product engineering for US-based clients in a fully remote, high-autonomy consultancy.

    • Built React front ends and REST APIs, owning features from spec through implementation, review and deploy.
    • Test-driven development as the default, not an afterthought — a habit that still shapes how I ship AI systems.
    • Worked directly with client stakeholders across time zones: understand the real problem first, then build.

Skills

Grouped by where they actually get used, not by how impressive the list looks.

AI & LLM
PythonLangChainLangGraphCrewAIOpenAI Agents SDKMCPPyTorchFine-tuning (QLoRA)NLPPrompt engineeringMulti-agent orchestration
RAG & retrieval
Chunking strategiesEmbeddingsVector searchpgvectorPineconeQdrantHybrid searchReranking
LLMOps
Evals & regression gatesLangSmithLangfuseRAGASGuardrailsTracing & observabilityToken & cost control
Backend
PythonFastAPINode.jsTypeScriptPostgreSQLRESTGraphQLWebSocketsDistributed systemsData pipelines
Cloud & platform
AWS (Bedrock)Azure OpenAIVertex AIOracle Cloud (OCI)DockerKubernetesCI/CDnginx
Frontend
ReactNext.jsTypeScriptTailwind CSSFlutter

Certifications

The AI Engineer track courses plus Oracle Cloud certifications.

  • The Complete Agentic AI Engineering Course — Agent & MCP TrackUdemy · Jun 2025
  • AI Engineer Core Track: LLM Engineering, RAG, QLoRA, AgentsUdemy
  • AI Engineer Production Track: Deploy LLMs & Agents at ScaleUdemy
  • AI Coder: Complete Claude Code & Coding Agents CourseUdemy
  • AI Builder: Create Agents, Voice Agents & Automations in n8nUdemy
  • AI Leader: Generative AI & Agentic AI for Leaders & FoundersUdemy
  • Oracle Cloud Infrastructure 2025 Certified Developer ProfessionalOracle · Oct 2025
  • Oracle Cloud Infrastructure Foundations AssociateOracle · Sep 2025