HARDIK
SYSTEMS / SECURITY / ML

Hardik Ahlawat

RESEARCH. EXPERIMENT. MEASURE. BUILD.

I work at the intersection of cybersecurity, machine learning, and systems engineering — trusted execution environments, password security, network intrusion detection, vulnerability research, and autonomous network operations.

View ResearchView Projects
Research02
Featured systems02
GitHub, curated03
Based inNew Delhi
SCROLL01 / 09
01RESEARCH

Research

Research at the intersection of cybersecurity, machine learning, and trusted systems.

2026
Accepted — ETTIS 2026Springer / Scopus-indexed

Runtime Data Shadowing (RDS)

A TEE Approach for Secure Medical Data Sharing

A trusted-execution-environment approach for sharing medical data under role-aware access control, built on Intel SGX enclaves and authenticated encryption.

8.93 μsMean query latency
2.9×Overhead
112,046Throughputops/sec
15×Latency degradationbeyond 128 MB EPC boundary
Intel SGXTEEAES-128-GCM
Read case study
2026
Accepted — ICDSCNC 2026IEEE Xplore

Evaluating LSTM-Based Password Guessing

Under Argon2id Memory-Hard Constraints

Evaluates how effectively a character-level LSTM password guesser performs against credentials protected by Argon2id's memory-hard parameters, compared to a classical n-gram baseline.

2-layer LSTMArchitecture256 hidden units
2.29Validation loss
5.41×Efficiency vs. 3-gram Markov baseline
0.87%Attack match rate
666,401Credentials evaluated
LSTMArgon2idPython
Read case study
02PROJECTS

Featured Projects

Systems built end-to-end: from network simulation and telemetry to prediction and detection.

FLAGSHIP PROJECT · Systems / Networking / Applied ML

MPLS Predictive Copilot

Air-Gapped Autonomous NOC System

A 7-module air-gapped NOC pipeline combining Containerlab MPLS simulation, Prometheus telemetry, LightGBM impact prediction, NetworkX blast-radius analysis, and a local LLM copilot — zero outbound dependency.

7Pipeline modulesfully air-gapped
8Prediction featuresLightGBM time-to-impact model
ContainerlabPrometheusLightGBMNetworkXLocal LLMFastAPIStreamlit
View case studyRepository
SYSTEM TOPOLOGY
01MPLS Network
02Telemetry
03LightGBM Prediction
04Blast Radius Analysis
05Local LLM Copilot
06Operator
Machine Learning / Network Security

Deep-NIDS

AI-Powered Network Intrusion Detection System

A network intrusion detection system built on an LSTM autoencoder, using PCA-based feature reduction over the NSL-KDD dataset to flag anomalous traffic from reconstruction error.

85.4%Detection accuracy
41 → 10Feature reduction
12 msInference latency
LSTM AutoencoderNSL-KDDPCAPythonPyTorch
View case studyRepository
01Network Traffic
02Feature Extraction
03PCA
04LSTM Autoencoder
05Anomaly Detection
03MORE ON GITHUB

More on GitHub

Additional research and experiments, curated from public repositories.

AetherGraph

Security Research

Confidence-aware security-evidence graph fusion: a forensic failure analysis and correction (provenance, edge semantics, live-validated).

Python
0 0

Deep-Learning-for-Targeted-Threat-Mitigation

Machine Learning / Security

Detects spear-phishing emails with a hyperparameter-tuned LSTM classifier (TensorFlow/Keras, KerasTuner), optimized for high recall on enterprise threat data.

Python
1 0

AuthPrint

Machine Learning / Security

Identity-based moderation engine: flags AI-generated submissions by fingerprinting a creator's stylometric writing habits (character n-gram TF-IDF, stylistic drift) instead of analyzing text content.

Python
1 0
View full GitHub profile
04EXPERIENCE

Experience

  1. Jun 2026 – Present

    New Delhi — Hybrid

    Cybersecurity Analyst

    Peace Code

    Vulnerability assessment and risk prioritization for client systems.

    • Vulnerability assessments
    • CVSS-based risk prioritization
    • Remediation sequencing
    • Threat intelligence
    • Security architecture
  2. Jun 2026

    Remote

    Security Researcher

    IdeaPulley Bug Hunt

    Independent security research identifying authorization and business-logic flaws in a production API.

    • Critical API race condition
    • Rate-limit bypass
    • Credit-deduction logic bypass
    • Authorization bypass
    • Parameter tampering
    • Clerk auth-metadata manipulation
05EDUCATION

Education

Dual enrollment across two accredited institutions, pursued concurrently since 2024.

IIT Patna

B.Sc. (Hons) — Computer Science & Data Analytics

2024 – 2027
  • SGPA 10.0 / 10.0 — Semester 4
  • SGPA 9.0 / 10.0 — Semester 3

VIPS

B.Tech — Computer Science Engineering — Cybersecurity

2024 – 2028
06SKILLS

Skills

SECURITY & SYSTEMS

Penetration TestingVulnerability AssessmentCVSSAPI SecurityBurp SuiteNmapOWASP Top 10Intel SGXTEE Enclave IsolationAES-128-GCM

AI / ML / DATA

TensorFlowKerasPyTorchCUDALSTMLSTM AutoencodersPCANumPyPandasScikit-learn

ENGINEERING

PythonC/C++SQLGitDockerFastAPIJupyter
07ACHIEVEMENTS

Achievements

  • 9th Place — mythX Capture The FlagKIET University · 2026
  • Zonal Football & Kabaddi competitor
  • Winner — FIT-A-THON bodyweight fitness competition
08ABOUT

About

I work at the intersection of cybersecurity, machine learning, and systems engineering — trusted execution environments, password security, network intrusion detection, vulnerability research, and autonomous network operations.

RESEARCHER2 accepted publications

Trusted execution environments, password security under memory-hard KDFs.

ENGINEERTEE / ML / network security

Intel SGX enclaves, LSTM autoencoders, intrusion detection pipelines.

BUILDER2 featured systems

Air-gapped autonomous operations, edge-local intrusion detection.

09 / CONTACT

Let’s build something difficult.

Research, security, systems, or interesting engineering problems.

EmailGitHubLinkedInORCID
New Delhi, India09 / 09