Sameer Banchhor

Sameer Banchhor

AI Infrastructure & Systems Engineer • Backend Architecture • MLOps & High-Performance Data Platforms

I design and build fast, scalable, and cost-effective infrastructure for AI applications and data platforms. My work centers on high-performance backend architecture, large-scale distributed data pipelines, sub-15ms database query optimization, production model serving, and cloud deployment—turning complex AI workloads into resilient production systems.

"Designing high-throughput, low-latency infrastructure that powers reliable AI applications at scale."
Hugging Face → GitHub Profile → LinkedIn → Email Direct →
Voter Records Indexed 1,045,426 DuckDB 1.5 Engine (<15ms Latency)
Batch Records Processed 248,539 Scalable ETL & Inference Pipeline
API Latency SLA < 15ms Go 1.24 & Google Cloud Run
Neural Audio Scale 18M+ Chhattisgarhi VITS Inference Pipeline
Multi-Output Inference 5 Targets Concurrent Prediction & Early Warnings

AI Systems & Infrastructure Projects

Backend Architecture, Low-Latency Model Serving, Data Pipelines & Cloud Systems

Durg Voter Production REST API & Analytics Engine

Production System • Go 1.24 & DuckDB 1.5 Vector Engine
2026 • Production Live
High-performance, production-grade Go RESTful API & Embedded Analytics Dashboard serving 1,045,426 voter records and 1,513 polling station booths with query latency under 15ms.
Integrated DuckDB 1.5 vectorized analytics engine for instant electorate demography calculations across 1.04M+ records.
Built complete RESTful endpoint suite (EPIC profile lookup, multi-criteria voter search, polling stations, constituencies, OpenAPI 3.0 spec).
Developed dynamic AI transliteration toggle (English ↔ Hindi via Gemini API) with <50ms ultra-fast fallback search mode.
Architected enterprise middleware: Token Bucket Rate Limiting, Request ID context tracing, structured logging, panic recovery, & CORS control.
Live deployments on Google Cloud Run & Hugging Face Spaces with custom domain voter.sameerbanchhor.com.
Go 1.24 DuckDB 1.5 Google Cloud Run Gemini API RESTful API Docker OpenAPI 3.0
A production-grade RESTful backend & embedded dashboard serving voter demography and booth analytics for Durg region. Key Architecture & Technical Capabilities: 1. Vectorized Analytics (DuckDB 1.5): Direct integration with 1.04M voter database using CGO driver, achieving <15ms query execution speed. 2. Dual Cloud Architecture: Primary backend deployed on Google Cloud Run, mirror backend on Hugging Face Space, web frontend active at voter.sameerbanchhor.com. 3. AI Transliteration & Search: Dynamic English ↔ Hindi query transliteration powered by Gemini API, with fallback toggle (?use_ai=false) for sub-50ms searches. 4. Middleware Pipeline: Enterprise Token Bucket Rate Limiting, Request ID context tracing, panic recovery, & structured JSON logging.

DURG-EduAI — Large-Scale Data Pipeline & Multi-Task Inference Engine

Data Engineering • Scalable Machine Learning Pipelines
Aug 2025 – Feb 2026
High-throughput data engineering and ML pipeline designed for batch processing 248,539 academic records with concurrent multi-target inference.
Built scalable ETL feature pipelines in Python processing 248.5K structured records spanning a 10-year dataset.
Architected a 5-target parallel inference system delivering SGPA regression (R² ~ 0.996) and risk alerts in real-time.
Engineered a 6-signal proxy-labeling heuristic pipeline to detect academic risk without requiring longitudinal tracking.
Python ETL Pipelines XGBoost LightGBM Scikit-learn Data Architecture
DURG-EduAI is a high-throughput data processing and inference framework built to ingest, clean, and run predictions on 248,539 student examination records. Core System Capabilities: 1. Optimized ETL Pipelines: Cleans, structures, and validates 10 years of raw university records. 2. Parallel Multi-Output Inference Engine: Computes 5 distinct outputs per student in a single pass (SGPA, ATKT, Dropout Risk, Benchmarks, Alerts). 3. Scalable Feature Engineering: Memory-efficient feature extraction reducing dataset memory footprint while maximizing execution speed.

Chhattisgarhi Neural Audio Synthesis Pipeline (VITS Deep TTS)

Audio Infrastructure • Neural Synthesis Engine
2025 – 2026
End-to-end neural audio synthesis pipeline and model serving framework for Chhattisgarhi (18M+ native speakers).
Architected end-to-end neural synthesis pipeline from Devanagari text normalization to raw waveform audio generation.
Optimized VITS model inference pipeline for low-latency audio generation on constrained hardware.
Designed custom text preprocessing, phonetic tokenization, and streaming audio export scripts.
VITS Architecture PyTorch Model Serving Audio Processing HuggingFace Hub
End-to-end neural audio inference system built for Chhattisgarhi text-to-speech synthesis. Key Technical Architecture: - Deep VITS (Variational Inference with adversarial learning) neural network integration. - Custom Devanagari text normalization and phonetic processing pipeline to handle regional dialect nuances. - Streaming audio pipeline engineered for low-latency voice output in web and mobile applications.

Low-Latency Model Serving & Edge Conversational Infrastructure

Model Serving • Cloud APIs & Field Infrastructure
Mar 2025 – Jul 2025
Production model serving pipeline connecting fine-tuned LLM backends (Gemini API) to edge users in low-bandwidth rural environments.
Architected resilient API integration with fallback mechanisms to handle unstable connectivity in field deployments.
Fine-tuned and optimized prompt/context pipelines for real-time agricultural query response processing.
Deployed in production with active monitoring and continuous user feedback loops, recognized by local government.
Gemini API API Architecture Model Serving Cloud Deployment Resilient Systems
A robust, field-tested model serving framework built to deliver low-latency AI responses in low-resource environments. Key Highlights: - Cloud LLM serving (Gemini API) optimized for low latency and high availability. - Custom context caching and prompt engineering pipelines tuned for regional dialects. - Recognized and compensated by local government for community technology impact.

DU-Analysis — Data Engineering & Preprocessing Framework

Open Source • Data Pipelines
2025
High-throughput data extraction and transformation pipelines built to parse, clean, and structure unstructured academic datasets into optimized database formats.
Python ETL Pipelines Pandas Data Engineering

Infrastructure & Technical Stack

Backend Architecture, Systems, Databases & Model Serving
Backend Architecture & APIs
Go (Golang 1.24) RESTful APIs OpenAPI 3.0 Rate Limiting (Token Bucket) Structured Logging & Tracing Microservices
Databases & Vector Search
DuckDB 1.5 Vectorized Search Engine SQL Optimization Database Indexing Low-Latency Queries
Cloud Deployment & MLOps
Google Cloud Run Docker Containerization HuggingFace Spaces & Hub Model Serving Gemini API Git / GitHub Actions
Data Processing & ML Engines
Python ETL Pipelines Pandas & NumPy PyTorch VITS Neural Audio XGBoost / LightGBM Scikit-learn

Education & Background

Academic Degrees and Professional Certifications

M.Sc. in Computer Science

Hemchand Yadav University, Durg
8.21 SGPA (Sem III)
2024 – 2026

B.Sc. in Computer Science

Kalyan PG College, Bhilai
79%
2021 – 2023

Google Data Analytics Professional Certificate

Coursera
Completed
2026

IBM Data Science Professional Certificate

Coursera
Completed
2026