Data Engineering
PRORobust data pipelines, scalable ETL architectures, and cloud data warehouses.
Website Development
PROModern, responsive frontend architectures and high-performance web applications.
Data Science
PROAdvanced analytics, statistical modeling, and actionable business insights.
ML Pipelines
PROEnd-to-end machine learning orchestration and deterministic AI systems.
What we've built
6 production systems — shipped end-to-end with measurable outcomes.
- System Engineering / Generative AI
TTA Engine
99.4% success rate
Production pipeline reliability
Hybrid neuro-symbolic AI video generation pipeline. Converts academic topics into fully rendered, narrated educational videos at scale — deployed at Edza.ai.
PythonFastAPIGoogle Vertex AIRedisRead case study → - Computer Vision / Distributed Systems
Assessment Arena
200+ papers/hour
Automated grading throughput
AI-native evaluation infrastructure with multimodal answer ingestion, rubric-based semantic grading, and human-style annotated PDF feedback overlay.
PythonFastAPIMultimodal AIRedisRead case study → - Generative AI / Content Engineering
AI Notes Generator
80% cost reduction
Per-PDF generation cost saved
Cache-first pipeline converting curriculum data into CBSE-compliant PDF textbooks. 85% cache hit rate. Sub-second delivery for cached content.
PythonWeasyPrintGoogle Cloud StorageAsyncIORead case study → - System Design / Distributed Systems
E-News 2.0
100k+ concurrent users
Simultaneous connections handled
Distributed microservices media platform with HLS live streaming, algorithmic feed personalization, and real-time WebSocket breaking-news overlays.
GolangPythonRedisApache KafkaWebSocketsRead case study → - Data Science / Quantitative Finance
Time-Series Analysis
5,000+ data points
20 years of GOOGLE stock analyzed
Deterministic financial analytics pipeline: rolling volatility modeling, Sharpe ratio, max drawdown, and regime detection on OHLCV data.
PythonPandasNumPySciPyMatplotlibRead case study → - Machine Learning System Design
F1 2025 Analytics
92.4% R² score
Lap-time prediction accuracy
End-to-end F1 race strategy simulation platform. XGBoost regression for lap prediction, SHAP explainability, interactive pit-window scenario engine.
PythonXGBoostSHAPFastAPINext.jsRead case study →
Products
A small set of projects—built end-to-end with real engineering.
About us
We're two engineers building product-grade systems with great UX.
- MP
Munish Pathania
Software Engineer
ViewMachine learning with curiosity.
Next.js + TypeScriptPythonData pipelines - RB
Rishabh Bhartiya
Machine Learning Engineer
ViewBuilding production-grade AI systems with strong evaluation and data foundations.
LLM EvaluationEducational AI SystemsData-Driven Optimization