Syed Hashir Mahmood
Open to work2 years experienceSenior Software Engineer
AI Engineer specializing in LLM-powered systems, RAG pipelines, and distributed microservices.
Standing Rank
Rank Not Available
Developer Badges
No badges earned yet
Skills & Technologies
53 skillsWork Experience
Senior Software Engineer
KPIT Technologies
- Architected an internal developer platform (IDP) adopted company-wide, orchestrating 500+ daily CI/CD pipelines via event-driven Kafka microservices, decoupling execution from reporting with async state monitoring for real-time observability.
- Designed a distributed concurrency-control algorithm managing a 10-job concurrent execution pool for contended license allocation, eliminating resource starvation and guaranteeing zero scheduling conflicts under peak load.
- Drove AWS cost down 30% and simulation execution time 20% by automating full EC2 fleet lifecycle management (provision, scale, reclaim) with idle- detection controls and MinIO-backed artifact versioning.
- Hardened platform security by building centralized IAM with SSO and RBAC via Keycloak (OAuth2/OIDC), securing service-to-service auth across distributed microservices.
Projects
AskDocs \u2014 Agentic RAG Document Intelligence
View Project- Built a document Q&A RAG service that parses uploads (Apache Tika), chunks and embeds them locally via Ollama, and answers questions over pgvector similarity search (HNSW) with source citations grounded in retrieved text.
- Added an autonomous agent that reasons over Model Context Protocol (MCP) tools, chaining multiple retrievals to answer multi-part questions a single retrieval pass cannot, with fail-soft tool handling so hallucinated arguments degrade gracefully instead of crashing requests.
- Built a 25-question evaluation harness over a 125-chunk corpus measuring retrieval quality: 92% recall@5 at 44ms median latency, verified end-to-end against real Postgres using Testcontainers.
AI-First DevOps Platform
View Project- Integrated an LLM (Claude API) at three pipeline stages: Terraform plan security review, dynamic Kubernetes resource configuration from plain-English goals, and post-deploy health analysis with automatic rollback on degraded deployments.
- Deployed a containerized full-stack app to GCP (GKE + Cloud SQL) via a GitHub Actions CI/CD pipeline with zero-downtime rolling deployments and auto- rollback on failure.
- Wrote cloud-agnostic Terraform modules enabling identical infrastructure across AWS (EKS + RDS) and GCP (GKE + Cloud SQL) by changing a single configuration variable.
Incident Response Agent
View Project- Developed a LangGraph agent that investigates incidents through an iterative reasoning loop, dynamically selecting tools based on prior findings and self- assessing evidence sufficiency with a hard iteration cap as a safety backstop; added persistent memory (PostgreSQL) to recall past incidents.
- Implemented human-in-the-loop approval via LangGraph's interrupt/resume, so the agent escalates on low confidence instead of acting autonomously, pausing for human sign-off before any remediation. Runs on local LLMs (Ollama), zero API cost.
- Designed a 14-case adversarial eval suite with a custom grader; used it to fix a confidence-calibration bug (30% to 70%), then extended to a two-agent architecture (supervisor for judgment, investigator for evidence gathering) to isolate reasoning bias, improving pass rate 71% to 79%.
Distributed Rate Limiter Service
View Project- Built a multi-tenant rate limiter serving 6,200+ req/sec at sub-5ms p99 latency, using atomic Redis Lua scripts to eliminate read-modify-write races under concurrent load.
- Implemented three interchangeable algorithms (token bucket, sliding window log, fixed window) behind a strategy interface, with per-tenant algorithm and limit changes applied at runtime through an admin REST API, no redeploys needed
- Validated correctness across 3 nginx-balanced instances under 1,000-request bursts with zero state inconsistencies, backed by k6 benchmarks and JUnit/Testcontainers tests covering edge cases like Redis failure.
Leaderboard Standings
Leaderboard Position Pending
Global test scores, peer standing percentiles, and algorithm leaderboard ranks are updated dynamically.
Assessment Highlights
Assessments Not Completed
Coding evaluations, system assessment results, and conceptual score badges will appear here after taking a test.
AI Collaboration Score
AI Collaboration Score Pending
Developer coding behavior, assistant cooperation, and AI pair-programming indicators are evaluated during live coding sessions.
Role Compatibility Profile
Role Compatibility Analysis Pending
Custom matching reports, candidate role compatibility percentiles, and core engineer strength profiles are processed once conceptual code screenings are complete.
Achievements
GATE CSE 2022 Rank
AIR 905 in GATE CSE 2022.
About Details
Professional Bio
AI Engineer with production software background at KPIT and Amazon, now focused on building LLM-powered systems: RAG pipelines, autonomous agents, and MCP-based tool use. Experienced in distributed systems, event-driven microservices, and CI/CD platform development.
M.Tech in Computer Science and Engineering
International Institute of Information Technology, Bangalore (2022 - 2024)
B.Tech in Information Technology
Bharati Vidyapeeth (Deemed to be University) College of Engineering, Pune (2017 - 2021)