OP / 01VERBS-FIRST
01 ·
Build
agentic systems.
Tool-use, memory, multi-agent orchestration.
- LangChain
- LangGraph
- Deep Agents
- Ollama
- MCP
Founder · Savira AI·AI / LLM Engineer·Open to roles · IN + NP + remote
Sajid Miya›
[ MISSION ]The pipeline below is what I shipped last month.
FIG.01 / Retrieval-augmented agent flow5 STAGES · 1 OUTPUT
STG/0101/05
Source
Documents · web · policies
STG/0202/05
Engine
Parse · structure · version
STG/0303/05
Recall
Hybrid search · rerank · filter
STG/0404/05
Agents
Intent · compare · compliance
STG/0505/05
Answer
Citations · confidence
STAT / 01
3×
AI products shipped, end-to-end
STAT / 02
1×
AI studio · Savira, private alpha
STAT / 03
0.754
Sentence fluency Pearson (HiPPA)
[01·ABOUT]
▸Operator
// the short versionFILE001.ABOUT
STATUSACTIVE
PARAS3
LANGEN-IN
01I build AI systems that survive contact with production. Most of my work is around agentic systems, retrieval, and the boring infrastructure that makes LLMs usable — versioning, permissions, evaluation, observability.
02Stack: Python, FastAPI, the LangChain / LangGraph ecosystem. I have shipped RAG pipelines over FAISS, ChromaDB, and Pinecone, and wired enough APIs to know that the integration is the product.
03I write runbooks before code. If a system cannot be debugged at 2am, it does not ship. If a model change cannot be evaluated, it does not deploy.
[02·CAPABILITIES·4 GROUPS]
▸What I do
// verbs firstOP / 02VERBS-FIRST
02 ·
Wire
retrieval pipelines.
Hybrid search, reranking, contextual chunking.
- Agentic RAG
- GraphRAG
- FAISS
- ChromaDB
- Pinecone
OP / 03VERBS-FIRST
03 ·
Engineer
LLM backends.
Typed APIs, queues, observability, deployable.
- FastAPI
- Python
- PostgreSQL
- Redis
- Docker
OP / 04VERBS-FIRST
04 ·
Evaluate
model changes.
Eval harness gates every prompt change in CI.
- PyTorch
- TensorFlow
- Pandas
- Evals
- AWS
STACK
Python · FastAPI · LangGraph
VECTORS
FAISS · Chroma · Pinecone
DEPLOY
Docker · AWS · PostgreSQL
PRACTICE
Runbook before code
[03·EXPERIENCE·4 ROLES]
▸Field record
// reverse chronologicalEXP/01NOW
Aug 2026 — Present
Remote
Savira AI
// Founder & EngineerEnterprise AI infrastructure studio. Currently shipping Savira Cortex — a knowledge OS for internal documents.
- Designed the architecture for a multi-tenant knowledge OS: document ingestion, version-aware retrieval, multi-agent reasoning, RBAC enforced before retrieval.
- Defined and shipped a four-product surface: Cortex (knowledge OS), NexHire (hiring copilot), Agents (tool-using workflows), Voice (planned).
- Founded 10 August 2026. Public website and private alpha design-partner program live at savira.sajidmiya.tech.
FastAPILangGraphPostgreSQLVector DBMulti-Agent
EXP/02
2024
Independent
NexHire
// Creator · AI-Native Hiring PlatformEnd-to-end recruitment platform — job posting through offer — with LLM-driven screening, matching, scheduling, and email.
- Automated resume screening, candidate matching, interview scheduling, and email generation using LLMs, RAG, and agentic workflows.
- Integrated Google Calendar, Gmail, Slack, and Stripe for scheduling, communication, team notifications, and subscription billing.
- Stack: React, FastAPI, PostgreSQL, Redis, Docker, JWT.
EXP/03
2023
Independent
SynapseLearn
// Creator · Conversational AI TutorDesktop AI learning platform with multi-user auth, RAG over PDFs and web, and persistent memory threads.
- Built secure authentication and conversation management for multiple concurrent users.
- Designed a RAG pipeline over PDFs and web resources using FAISS + ChromaDB with contextual retrieval.
- Shipped memory-aware chat — persistent threads, session tracking, conversation merging — plus an admin dashboard.
EXP/04
2025
Independent
Portfolio Agent
// Creator · Local-First AI AssistantPrivacy-first AI assistant. Zero external API dependency. Fully local multi-modal.
- Built with FastAPI, LangGraph, Ollama, and SQLite — real-time streaming chat with no cloud calls.
- Designed persistent SQLite + JSON long-term memory for coherent context across sessions.
- Added vision input, tool calling, and dynamic model orchestration for fully local multi-modal deployment.
FastAPILangGraphOllamaSQLite
[04·RESEARCH·PEARSON · r]
▸HiPPA
// SpeechOcean762 benchmarkVENUESpeechOcean762 benchmark
TASKPronunciation Assessment
ARCHHierarchical MT-Transformer
SCOPEBeyond phoneme baseline
HiPPA — Hierarchical Multi-Task Transformer for Pronunciation Assessment.
Independently built a hierarchical multi-task transformer pipeline for pronunciation assessment on SpeechOcean762, extending the benchmark's coverage beyond its phoneme-only baseline.
ESTABLISHED TRANSFORMER BLOCKS (WavLM, cross-attention, multi-task learning, CTC) APPLIED TO A BENCHMARK WHOSE OFFICIAL BASELINE ONLY COVERS PHONEME-LEVEL SCORING.
Read paper (PDF)→METRIC / 01 · SENTENCE FLUENCY PEARSONLIVE
0.000
+67% relativevs baseline 0.450 (phoneme-only)
#METRICPEARSONDELTA
02Sentence fluency Pearson0.754▲ NEW
03Sentence prosody Pearson0.737▲ NEW
04Sentence total Pearson0.697▲ NEW
The trade-off: a small drop on phoneme-level scoring (0.450 → 0.396), in exchange for three new metrics the benchmark did not previously support — word total, sentence prosody, sentence fluency.
[05·CREDENTIALS·3 STAGES]
▸Education
// and awardsSECTION / 01 · ACADEMIC3 ENTRIES
- EDU/012023 — 2027
Thapar Institute of Engineering and Technology
B.E. Computer Science and Engineering
SCORE8.68 / 10 CGPAPatiala, Punjab
- EDU/022021 — 2023
St. Xavier's College, Maitighar
School Leaving Certificate (SLC)
SCORE3.66 / 4 GPAKathmandu, Nepal
- EDU/032021
The Old Capital Secondary School
Secondary Education Examination (SEE)
SCORE3.95 / 4 GPARaniban, Gorkha, Nepal
SECTION / 02 · AWARDS1 ENTRY
AWD/01
COMPEX Scholarship — Government of India (EdCIL)
2023 — Present
Competitively awarded merit scholarship for Nepalese students studying in India. Covers full tuition and hostel at Thapar.