Skip to content
View salimkt's full-sized avatar
🎯
Focusing
🎯
Focusing

Block or report salimkt

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
salimkt/README.md

Muhammed Salim K T

Software Engineer (AI/ML): LLM systems, RAG, agent orchestration, speech & vision
4+ years shipping production AI · Kochi, India

Portfolio · LinkedIn · mohdsalimkt@gmail.com


I build production LLM systems end to end: agent orchestration, retrieval, speech pipelines, model serving, and the evaluation that keeps them honest. Most of it runs on-premise and air-gapped inside enterprise customers' own infrastructure.

At a glance

  • 12B → 32B-class: my multi-agent engine lets a 12B model do 32B-class work, moving each deployment from a GPU cluster to a single node
  • 10M+ embeddings in production RAG, serving 10k+ queries/day
  • +45% answer accuracy over the bare-LLM baseline, 1.8 s average inference latency, 99.7% uptime
  • 94% speech-to-text accuracy and 92% live-caption accuracy in production
  • 50k+ concurrent users on real-time WebSocket / REST services
  • 8 Kubernetes environments, including two GPU/DGX clusters

What I have built at work

Multi-agent tool-calling engine. bm-agent-engine replaced the CrewAI-based orchestration of an on-premise enterprise LLM platform and now serves all its agent traffic; I was the largest contributor to its ~49k lines. A 73-toolset registry, ReAct agents with tiered escalation, and a FAST / SMALL / LARGE / THINKING router over vLLM and Ollama. Because the SSE event contract was kept exactly, the UI cut over with a single config change.

Enterprise RAG. Docling + RapidOCR ingestion, Sentence-Transformers embeddings, a Cassandra vector schema, and hybrid keyword + semantic retrieval with re-ranking over 10M+ embeddings. Quantization, batching and caching brought average inference to 1.8 s; model failover and warm caching hold 99.7% uptime.

Speech AI in production. A transcription service for a US hosted-telephony provider's voicemail and call recordings: NVIDIA Canary and Whisper on vLLM with automatic failover, Redis priority queues and Celery workers, recordings up to 10 hours, and speaker diarization. Also: Whisper STT (94%), Tacotron2 TTS and live captions (92%) in a video-calling product.

Vision and documents. PaddleOCR-VL and Qwen-VL served through Transformers, a pluggable OCR layer feeding RAG, and LayoutLMv3 fine-tuned to 91% F1 on 100K+ documents a month.

Selected public work

Project What it is
DataRat · live demo AI data analyst. Connect 11 kinds of data source and get KPIs, anomalies and Holt-Winters forecasts; a Planner → SQL → Analyst → Narrator agent pipeline answers plain-English questions, with a rules engine as fallback and a read-only SQL guard. FastAPI, DuckDB, React.
3xDezine · live demo Home design and costing. 2D floor plan to WebGPU 3D walkthrough with a live material estimate and GST-ready quotation; 300+ unit tests and 136 pgTAP database assertions. TypeScript, Fastify, Supabase, Kotlin.
small-model-agents Reference implementation of reliable agents on small models: schema-validated tool calling, tiered routing, evidence-based escalation, an MCP server, and an eval harness whose CI gate blocks regressions. Python stdlib.
web-share · live demo Peer-to-peer WebRTC file sharing as a PWA, with end-to-end-encrypted MQTT signalling or offline QR pairing. No server, no accounts.

DataRat executive dashboard with KPI cards, a Holt-Winters forecast and generated key findings 3xDezine studio with floor plan editor, WebGPU 3D view and live estimate

Open-source contributions

Project Contribution Status
Celery Log at INFO when a prefork child is recycled In review
PaddleOCR MCP server and doc2md fixes: #18416, #18417, #18418 In review
vLLM Design for multi-channel audio reduction Proposed
silero-vad Asymmetric speech padding Proposed

Stack

Python FastAPI Django Celery vLLM Ollama Hugging Face PyTorch LangChain CrewAI MCP Cassandra pgvector Qdrant FAISS PostgreSQL DuckDB Redis Kafka RabbitMQ Whisper NVIDIA Canary Tacotron2 LayoutLMv3 PaddleOCR-VL Qwen-VL Kubernetes Docker GCP GitLab CI GitHub Actions TypeScript React React Native three.js Kotlin


Software engineer (AI/ML) at Netstratum Technologies since June 2022. Contact: mohdsalimkt@gmail.com

Pinned Loading

  1. 3xdezine 3xdezine Public

    Home design and costing: 2D floor plan to photorealistic WebGPU 3D, with live material costing and GST-ready quotations. React Three Fiber, Fastify/Postgres, Supabase, Kotlin/Filament.

    TypeScript

  2. DataRat DataRat Public

    AI decision intelligence for business leaders: 11 data connectors, auto KPIs, anomaly detection, forecasting, and multi-agent natural-language analytics (FastAPI + DuckDB + React)

    Python 1

  3. salim-portfolio salim-portfolio Public

    Portfolio: Muhammed Salim K T, AI/ML engineer (voice AI, agents, RAG)

    HTML

  4. small-model-agents small-model-agents Public

    Reliable agent tool calling on small models: schema-validated decisions, tiered routing with evidence-based escalation, an MCP server and a CI eval gate. Python stdlib only.

    Python

  5. web-share web-share Public

    Local network file sharing via WebRTC — no cloud, no accounts, works offline on your WiFi

    JavaScript