Senior Research Engineer, Threat Intelligence

SecurityScorecard•Colorado Springs, CO

15h•$140,000 - $150,000

About The Position

You'll join STRIKE, SecurityScorecard's Threat Intelligence team, as the engineering counterpart to research. STRIKE runs several research motions in parallel, each on its own clock: rapid response to active events, longer product-tied work, and standards-anchored research on a quarterly cadence. The path from a finding to a shipped detection or feed gets reinvented every time. That's the problem this role is here to solve. You'll work directly with the senior technical leader who owns STRIKE's R&D direction, and report to the Head of Threat Research for people management. Technical direction comes from R&D leadership; you own delivery. You'll take a research artifact (a malware finding, an infrastructure cluster, a new indicator class, a behavioral pattern) and turn it into something the company can use without a second round of engineering: schemas, pipeline hooks, distribution feeds, detection rules, or platform APIs. This isn't a pure research role, and it isn't a pure platform role either. Researchers ideate, you ship.

Requirements

Bachelor's or Master's in Computer Science, Cybersecurity, or a related technical field. Self-taught practitioners with strong public work are welcome.
5 to 8 years in a hands-on engineering role with meaningful exposure to threat intelligence, security research, or detection engineering.
Prior experience building production systems that consume or emit threat intel data is required.
Python and TypeScript/Node at a production level
Relational and cache data stores, plus at least one streaming or batch data platform
Cloud infrastructure (AWS preferred), containers, and CI/CD pipelines
Working knowledge of STIX 2.1, TAXII 2.1, MISP, and MITRE ATT&CK, and how they work together in practice
Hands-on experience with YARA, Sigma, and STIX Patterning. Comfortable reading malware analysis output, parsing adversary infrastructure data, and writing detection logic that holds up under production load.
Shipped production systems that use language models, not just demos. That includes retrieval over a real corpus, structured output with schema validation, eval harnesses that catch regressions before users do, and a solid understanding of where models fail: recency, long-tail facts, numerical reasoning, and adversarial input or prompt injection. You can do the cost-per-task math for your workloads, and you can make the case when a smaller, tightly scaffolded model beats a larger one.
Approach model output with healthy skepticism by default. The bar for shipping a model-generated indicator or detection is higher than for shipping a regex, and you understand why and design accordingly.
Write code that ships, and you understand why researchers think the way they do. If you've only ever worked from a backlog handed down by a product manager, this probably isn't the right fit. If you've taken an idea sketched out in a chat message and turned it into a deployed pipeline before the next sprint began, that's the mode we're looking for.

Nice To Haves

Experience with policy-as-code or expression-language engines (CEL, OPA, or similar)
Published or co-authored security research (campaigns, vulnerabilities, adversary tracking)
Large-scale telemetry experience (Splunk, Kinesis, NetFlow, or equivalent)
Contributor or maintainer on open-source threat intel projects (MISP, OpenCTI, Sigma, STIX, ATT&CK)
Familiarity with quantitative risk frameworks such as FAIR
Familiarity with Golang at a production level

Responsibilities

Own the path from research output to production-ready artifact: a detection rule, a distributed feed, a scoring input, or a customer alert. Partner with adjacent teams to define clean handoff contracts, so new signals arrive downstream with the schema, value framing, and consumption pattern already defined.
Build and maintain STRIKE platform components across multiple services and runtimes, including distribution servers, sandbox orchestration, OSINT ingestion, federated sharing endpoints, agent runtimes, and rules engines that operate over standards-anchored predicates. Extend these systems without breaking the data contracts already in production.
Turn research into shipped detection content: YARA, Sigma, STIX patterns, behavioral indicators, and the pipelines that distribute them. Build correlation pipelines that link scan data, attack surface signals, vulnerability data, and adversary tracking into customer-facing intelligence.
Drive STIX 2.1 adoption as a unified output schema and TAXII 2.1 as a distribution standard. Define and govern schemas that hold up once they reach downstream teams.
Build the automation that removes commodity overhead from research work: indicator enrichment, report drafting, corpus correlation, feed normalization, and sandbox triage. Help move the team from analyst-driven, model-assisted workflows toward model-driven workflows with analyst review. The work that matters most here is often the unglamorous part: retrieval grounded in the team's own corpus so outputs cite sources rather than model priors, schema-constrained output so a generated indicator is a valid one, and eval harnesses that catch regressions before analysts do. Cost accounting, latency budgeting, prompt versioning, and output logging round out the infrastructure that makes a workflow safe to run unattended. You should have a clear sense of when a model is the wrong tool. A regex beats a model for known patterns; a SQL query beats a model for structured data. Knowing where that line sits, and respecting it, is part of the job.
Coordinate with engineering, measurement, and platform product teams so research actually lands in product. You'll often serve as the engineering voice translating between researchers, product managers, and platform engineers, and you may occasionally explain the work to customers, journalists, or executives.