Skip to content
View thisisvk45's full-sized avatar

Block or report thisisvk45

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Thisisvk45/README.md

Vikas Kumar

Co-founder of SViam, building in stealth. I build production AI systems: voice agents, retrieval (RAG) and ML pipelines that ship.

Before

  • Founding engineer at Lawroom AI, an AI legal research product used by 6,000 users
  • Built ML systems at Aditya Birla Group
  • MS, Wake Forest University

Technology

  • Languages: Python, TypeScript, SQL
  • AI: LLMs, RAG, voice agents (ElevenLabs, LiveKit), evaluation
  • Backend: FastAPI, Node.js, Postgres (Supabase), Redis
  • Frontend: Next.js, React
  • Infrastructure: AWS (Amplify, ECS, CloudFront), Docker

Elsewhere

helloviks.com · LinkedIn · Medium

Pinned Loading

  1. baml-diligence-eval baml-diligence-eval Public

    Controlled benchmark and adversarial stress test of BAML versus raw Pydantic for structured LLM output. 1800 calls, 95 percent confidence intervals, adversarial generation.

    Python

  2. hello-vikas hello-vikas Public

    Vikas Kumar | AI/ML Engineer | helloviks.com

    HTML

  3. InstantID InstantID Public

    Forked from instantX-research/InstantID

    Multi-image facial identity persistence on InstantID: confidence-weighted ArcFace fusion into a 512-dim identity vector, stored across sessions.

    Python

  4. Jobforge Jobforge Public

    Autonomous job search pipeline: Greenhouse/Lever/Ashby ingestion, Claude Haiku + Sonnet evaluation, RAG over ChromaDB, tailored CVs. 97/97 tests.

    Python

  5. prompt-doe prompt-doe Public

    Output-level entanglement diagnostics and Plackett-Burman screening for component-level prompt attribution. Code and data for the paper 'When Does Component Independence Hold?'

    Python

  6. prox-challenge prox-challenge Public

    Forked from prox-technologies/prox-challenge

    OmniPro Support: multimodal voice agent for the Vulcan OmniPro 220 welder. Multi-agent safety review, photo diagnosis. Built in 49 hours for Prox (YC F25).

    TypeScript