Abishek Giri

Work

Everything I've built, by area. The strongest thread is backend and distributed systems; the rest spans data, AI tooling, and full-stack. Each links to its source.

41 projects.

Distributed Systems & Infrastructure

  • ReleaseGate. Evidence infrastructure for software releases: deterministic, versioned policy decisions with signed, offline-verifiable proof of what was allowed and why, plus a customer-installable GitHub Action that signs and verifies deploys against a tenant. Python · FastAPI
  • Adaptive Release Controller. A self-adaptive CI/CD system that uses risk modeling and MAPE-K feedback loops to decide deploy, block, or rollback actions from real-time signals and historical outcomes. Python · CI/CD
  • Change Risk Predictor. An ML pipeline that scores the deployment risk of incoming code changes from commit history, CI signals, and ownership patterns. Python · Machine Learning
  • ShardDB. A fault-tolerant, strongly consistent distributed key-value store built from scratch in Go: Raft consensus, two-phase commit for transactions, automatic snapshotting, and linearizability verification under chaos. Go · Distributed Systems
  • MIDAS Ledger. A double-entry accounting ledger with strong correctness guarantees under extreme concurrency: atomic, idempotent transfers over async PostgreSQL with row-level locking and an immutable audit trail. Python · FastAPI
  • Job Orchestrator. A multi-tenant job orchestrator with race-free leasing, crash recovery, idempotent completion, retries with a dead-letter queue, and a transactional outbox. Python · PostgreSQL
  • Distributed KV Store. A distributed key-value store in FastAPI with leader-follower replication and synchronous replication across three nodes for strong consistency. Distributed Systems · Go
  • s3. A production-style Lakehouse architecture: ACID-compliant analytical storage on object storage with decoupled compute for batch, interactive SQL, and BI. Storage · AWS
  • Streaming Analytics Platform. A real-time streaming analytics platform that ingests, processes, aggregates, and stores high-volume events with event-time correctness and exactly-once guarantees. Flink · Kafka
  • Distributed KV Raft. A fault-tolerant, strongly consistent key-value store implemented from scratch in Go using Raft. Distributed Systems · Raft
  • PermissionDiff. A CI-native tool that shows how a code change alters effective authorization, replaying exact permission cases, enforcing invariants, and catching DENY to ALLOW regressions. Python · Security
  • MLSys 2026 Scheduler. A high-performance scheduler for the MLSys 2026 Graph Scheduling Challenge, optimizing subgraph fusion, granularity, and memory usage under a roofline model. Python · Scheduling

AI Research & Engineering

  • AgentLens. Local Python trace capture and evidence-based diagnosis for AI-agent failures. Python · TypeScript
  • LLM-CODEGEN. An exploration of how Groq's LLaMA-3. LLM · Generative AI
  • Large-Scale Bug Prediction. An end-to-end research project analyzing more than 700,000 source files across 10 major Python repositories (Django, NumPy, Pandas, Flask) for bug prediction. Big Data · Research
  • AI Recommender. An intelligent recommendation engine for personalized content. Recommender Systems · Python
  • AI in Software Engineering. Research on applying AI techniques to software-engineering problems. Research · AI
  • Research Paper. An academic research paper repository. Research · LaTeX
  • VLM Failure Modes. An investigation of whether Vision-Language Model uncertainty (such as entropy) actually rises on noisy, ambiguous, or corrupted inputs in safety-critical settings. VLM · AI Safety
  • VLM Adversarial Defense. An investigation of the adversarial robustness of Vision-Language Models (LLaVA-1. VLM · Adversarial ML
  • Amnesic VLM Defense. An investigation of why adversarial hallucinations occur in VLMs, using entropy analysis, attention visualization, and focal ablation to show they are confident, robust, and distributed rather than localized. VLM · Defense
  • AI Impact on Jobs. An analysis of AI's impact on the job market. Data Analysis · Research
  • Sentiment Steering GPT. Activation steering on GPT-2 Medium to identify and manipulate a moral direction, inverting outputs on negative concepts without retraining. LLM · Python
  • Criteo CTR Pipeline. A real-time data pipeline (Kafka, Spark Structured Streaming, Redis, FastAPI) processing simulated Criteo ad-click logs. ML · CTR

Full Stack & Web

  • Travel Buddies. An app to connect with people to travel around the world. React · Node.js
  • StorePro Fullstack. A full-stack store-operations app with inventory tracking, product management, authentication, and real-time updates. React · Node.js
  • NYC Taxi Pipeline. A scalable NYC Taxi data pipeline featuring Spark processing, ML predictions, and an interactive dashboard. Streamlit · Python
  • HighPerf Parquet. A from-scratch Parquet writer with a SIMD-accelerated encoding engine, adaptive encoding, and a GPU-aware compression pipeline, targeting 3 to 5x faster writes than traditional writers. C++ · C
  • Data Watermarking Engine. A framework for embedding and detecting invisible watermarks in tabular data (CSV, Parquet). Python · Security
  • PlanetStore Dashboard. A distributed object-storage engine inspired by S3: erasure coding, deduplication, versioning, multi-region replication, quotas, auth, rate-limiting, and a React admin dashboard. React · Dashboard
  • Real-Time Chat. A high-performance chat application built on WebSockets. WebSockets · Socket.IO
  • Pathfinding Visualizer. Interactive visualization of complex pathfinding algorithms. Algorithms · Visualization
  • Dementia Data Processor. A data-processing pipeline for dementia research analysis. Healthcare · Data Analysis
  • Python Search Engine. A custom search engine with indexing and retrieval built in Python. Information Retrieval · Algorithms
  • BasketMind. A basketball analytics tool for strategic insights. Sports Analytics · Data Science
  • Navigation System. Route-planning and navigation logic built on graph algorithms. Graph Theory · Routing
  • WeatherBuddy. A real-time weather forecasting application. API Integration · Frontend
  • Bayes Gaussian Plot. A visualization of Bayesian Gaussian processes. Bayesian · Math
  • Statistics Calculator. A utility for common statistical computations. Statistics · Math
  • Portfolio Website. This personal website. Flask · Tailwind CSS
  • Adaptive Layout Engine. A workload-aware storage optimizer that switches between date-partitioning and Z-order indexing to cut query latency in resource-constrained Spark. CSS · Layout