Building intelligent
systems that matter
Software engineer and ML researcher specializing in agentic AI, computer vision, and real-time systems. Currently exploring the intersection of healthcare and machine learning.
Personal Projects
A comprehensive collection of 12 machine learning and 30 quantitative finance projects demonstrating end-to-end implementation from research to production.
Image Classification
Deep learning image classification using CIFAR-10/ImageNet with ViT, EfficientNet, and ConvNeXt architectures.
Object Detection
Real-time object detection with YOLOv8, DETR, and Detectron2 achieving 30+ FPS inference.
Instance Segmentation
Pixel-level object segmentation using Mask R-CNN, SAM, and panoptic segmentation on Cityscapes.
Text Classification
NLP sentiment analysis with BERT and RoBERTa fine-tuning achieving 95%+ accuracy on IMDb.
Text Generation
Language model fine-tuning with GPT-2 and LoRA achieving perplexity under 25 on WikiText.
Machine Translation
Neural machine translation with MarianMT and M2M-100 for multilingual translation.
Speech Emotion Recognition
Audio-based emotion classification with Wav2Vec2 achieving 87%+ accuracy on RAVDESS.
Automatic Speech Recognition
End-to-end ASR with Whisper and Wav2Vec2 achieving WER under 5% on LibriSpeech.
Recommender System
Neural collaborative filtering with GNN achieving RMSE under 0.82 on MovieLens.
Time Series Forecasting
Multi-horizon forecasting with Temporal Fusion Transformer and N-BEATS on M4/ETT datasets.
Anomaly Detection
Unsupervised anomaly detection with autoencoders and isolation forests for industrial IoT.
Multimodal Fusion
Vision-audio-text fusion with cross-attention achieving 15-30% accuracy improvement.
Adaptive Market Making
C++ market making engine with Python RL agent for dynamic quote skewing and inventory management.
Order Book Imbalance Scalper
Sub-second trading on micro-alpha signals from order flow imbalance and queue depth analysis.
Queue Position Modeling
C++ queue position estimation with dynamic cancel/replace for top-of-book priority.
Cross-Exchange Arbitrage
Latency-aware multi-venue arbitrage accounting for fees, rebates, and cancel risk.
Short-Horizon Trade Imbalance
Predict next N ticks using OFI, trade signs, and microprice with sub-millisecond signals.
Iceberg Detection & Sniper
Detect hidden liquidity via partial fills and refill patterns for optimal execution.
Latency Arbitrage Simulator
Educational simulator for studying stale quote exploitation with controllable jitter.
Smart Order Router (SOR)
Rebate-aware cost model minimizing impact, fees, and opportunity cost across venues.
RL Market Maker
PPO/DQN agents on C++ LOB environment with Gym-style interface via pybind11.
Factor Research Toolkit
Cross-sectional equity/crypto factor research with point-in-time data handling and Alphalens validation.
Event-Driven Backtester
Comprehensive backtesting framework with vectorized and event-driven engines and walk-forward CV.
Statistical Arbitrage Framework
Cointegration-based pair/cluster trading with Kalman filter hedge ratios and regime filtering.
Options Volatility Surface
Arbitrage-free vol surface construction with SVI/SABR calibration and Greeks computation.
Limit Order Book Simulator
High-performance C++ LOB engine with Poisson/Hawkes arrivals and Python bindings.
Execution Algorithms
VWAP, TWAP, POV, and IS algorithms with market impact modeling and TCA analytics.
Real-Time Feed Handler
Low-latency C++ market data handler with lock-free structures and NUMA optimization.
Momentum / Trend-Following
EMA/SMA crossovers, breakout detection, and multi-timeframe momentum strategies.
Mean Reversion Strategy
Z-score reversion, Bollinger Bands, and RSI-based mean reversion with pairs trading.
Statistical Arbitrage
Engle-Granger and Johansen cointegration with rolling hedge ratios for pairs/clusters.
Momentum vs Value Long-Short
Cross-sectional factor strategy with sector/market neutralization and risk-parity sizing.
Options Strategy
Earnings straddles, systematic covered calls, and IV-RV spread trading.
Execution & TCA
C++ execution algorithms (VWAP/TWAP/POV) with impact models and slippage analysis.
Machine Learning Strategy
LSTM/Transformer forecasting with embargoed cross-validation and calibration.
Regime Detection & Allocation
HMM/Markov-switching regime detection with dynamic momentum/mean-reversion allocation.
Portfolio Construction & Risk
EWMA/Ledoit-Wolf covariance estimation with Kelly sizing and risk constraints.
Advanced Regime Detection System
Sophisticated HMM/Markov-switching regime detection with HDBSCAN clustering and meta-policy allocation.
LSTM/Transformer Forecasting System
Advanced time series forecasting with embargoed cross-validation, calibration, and Kelly sizing.
Gym LOB RL Market Maker
OpenAI Gym limit order book environment with DQN/PPO/SAC agents for market making.
Hawkes Process LOB Engine
C++ discrete-event LOB simulator with Hawkes process arrivals and Python bindings.
Ultra-Low Latency Feed Handler
C++ market data feed handler with lock-free queues, FIX encoding, and p50/p99 latency tracking.
Get In Touch
Interested in collaborating, have a job opportunity, or just want to say hello? I'd love to hear from you.
Or reach out directly via email
anishguntreddi@gmail.com